zint/backend/qr.c

2809 lines
97 KiB
C
Raw Permalink Normal View History

/* qr.c Handles QR Code, Micro QR Code, UPNQR and rMQR */
/*
2016-02-17 23:52:48 +13:00
libzint - the open source barcode library
Copyright (C) 2009-2024 Robin Stuart <rstuart114@gmail.com>
2016-02-17 23:52:48 +13:00
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
1. Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
2016-02-17 23:52:48 +13:00
2. Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
2016-02-17 23:52:48 +13:00
3. Neither the name of the project nor the names of its contributors
may be used to endorse or promote products derived from this software
without specific prior written permission.
2016-02-17 23:52:48 +13:00
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
2016-02-17 23:52:48 +13:00
SUCH DAMAGE.
*/
/* SPDX-License-Identifier: BSD-3-Clause */
2016-02-17 23:52:48 +13:00
#include <assert.h>
#include <math.h>
2016-02-17 23:52:48 +13:00
#include <stdio.h>
#include "common.h"
#include "eci.h"
2016-02-17 23:52:48 +13:00
#include "qr.h"
#include "reedsol.h"
#define QR_ALPHA (IS_NUM_F | IS_UPR_F | IS_SPC_F | IS_AST_F | IS_PLS_F | IS_MNS_F | IS_SIL_F | IS_CLI_F)
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
#define QR_LEVEL_L 0
#define QR_LEVEL_M 1
#define QR_LEVEL_Q 2
#define QR_LEVEL_H 3
static const char qr_ecc_level_names[] = { 'L', 'M', 'Q', 'H' };
#define QR_PERCENT 38 /* Alphanumeric mode % */
#define RMQR_VERSION 41
#define MICROQR_VERSION 73
/* Returns true if input glyph is in the Alphanumeric set or is GS1 FNC1 */
static int qr_is_alpha(const unsigned int glyph, const int gs1) {
if (is_chr(QR_ALPHA, glyph)) {
return 1;
}
if (gs1 && glyph == '\x1D') {
return 1;
}
return 0;
2016-02-17 23:52:48 +13:00
}
2019-12-17 06:31:52 +13:00
/* Bits multiplied by this for costs, so as to be whole integer divisible by 2 and 3 */
#define QR_MULT 6
2016-02-17 23:52:48 +13:00
/* Whether in numeric or not. If in numeric, *p_end is set to position after numeric, and *p_cost is set to
* per-numeric cost */
static int qr_in_numeric(const unsigned int ddata[], const int length, const int in_posn,
unsigned int *p_end, unsigned int *p_cost) {
int i, digit_cnt;
2019-12-17 06:31:52 +13:00
if (in_posn < (int) *p_end) {
2019-12-17 06:31:52 +13:00
return 1;
}
/* Attempt to calculate the average 'cost' of using numeric mode in number of bits (times QR_MULT) */
for (i = in_posn; i < length && i < in_posn + 3 && z_isdigit(ddata[i]); i++);
2019-12-17 06:31:52 +13:00
digit_cnt = i - in_posn;
2019-12-17 06:31:52 +13:00
if (digit_cnt == 0) {
*p_end = 0;
return 0;
}
*p_end = i;
*p_cost = digit_cnt == 1
? 24 /* 4 * QR_MULT */ : digit_cnt == 2 ? 21 /* (7 / 2) * QR_MULT */ : 20 /* (10 / 3) * QR_MULT) */;
2019-12-17 06:31:52 +13:00
return 1;
}
/* Whether in alpha or not. If in alpha, *p_end is set to position after alpha, and *p_cost is set to per-alpha cost.
* For GS1, *p_pcent set if 2nd char percent */
static int qr_in_alpha(const unsigned int ddata[], const int length, const int in_posn,
unsigned int *p_end, unsigned int *p_cost, unsigned int *p_pcent, unsigned int *p_pccnt,
unsigned int gs1) {
const int last = in_posn + 1 == length;
2019-12-17 06:31:52 +13:00
int two_alphas;
/* Attempt to calculate the average 'cost' of using alphanumeric mode in number of bits (times QR_MULT) */
if (in_posn < (int) *p_end) {
if (gs1) {
if (*p_pcent) {
/* Previous 2nd char was a percent, so allow for second half of doubled-up percent here */
two_alphas = !last && qr_is_alpha(ddata[in_posn + 1], gs1);
/* Uneven percents means this will fit evenly in alpha pair */
*p_cost = two_alphas || !(*p_pccnt & 1) ? 66 /* 11 * QR_MULT */ : 69 /* (11 / 2 + 6) * QR_MULT */;
*p_pcent = 0;
} else {
/* As above, uneven percents means will fit in alpha pair */
*p_cost = !last || !(*p_pccnt & 1) ? 33 /* (11 / 2) * QR_MULT */ : 36 /* 6 * QR_MULT */;
}
} else {
*p_cost = !last ? 33 /* (11 / 2) * QR_MULT */ : 36 /* 6 * QR_MULT */;
}
2019-12-17 06:31:52 +13:00
return 1;
}
2016-02-17 23:52:48 +13:00
if (!qr_is_alpha(ddata[in_posn], gs1)) {
2019-12-17 06:31:52 +13:00
*p_end = 0;
*p_pcent = 0;
*p_pccnt = 0;
2019-12-17 06:31:52 +13:00
return 0;
}
two_alphas = !last && qr_is_alpha(ddata[in_posn + 1], gs1);
if (gs1 && ddata[in_posn] == '%') { /* Must double-up so counts as 2 chars */
*p_end = in_posn + 1;
2019-12-17 06:31:52 +13:00
*p_pcent = 0;
(*p_pccnt)++;
/* Uneven percents means will fit in alpha pair */
*p_cost = two_alphas || !(*p_pccnt & 1) ? 66 /* 11 * QR_MULT */ : 69 /* (11 / 2 + 6) * QR_MULT */;
2019-12-17 06:31:52 +13:00
return 1;
}
*p_end = two_alphas ? in_posn + 2 : in_posn + 1;
if (gs1) {
*p_pcent = two_alphas && ddata[in_posn + 1] == '%'; /* 2nd char is percent */
*p_pccnt += *p_pcent;
/* Uneven percents means will fit in alpha pair */
*p_cost = two_alphas || !(*p_pccnt & 1) ? 33 /* (11 / 2) * QR_MULT */ : 36 /* 6 * QR_MULT */;
} else {
*p_cost = two_alphas ? 33 /* (11 / 2) * QR_MULT */ : 36 /* 6 * QR_MULT */;
}
2019-12-17 06:31:52 +13:00
return 1;
}
#if 0
#define QR_DEBUG_DEFINE_MODE /* For debugging costings */
#endif
/* Indexes into qr_mode_types array (and state array) */
2019-12-17 06:31:52 +13:00
#define QR_N 0 /* Numeric */
#define QR_A 1 /* Alphanumeric */
#define QR_B 2 /* Byte */
#define QR_K 3 /* Kanji */
static const char qr_mode_types[] = { 'N', 'A', 'B', 'K', '\0' }; /* Must be in same order as QR_N etc */
2019-12-17 06:31:52 +13:00
#define QR_NUM_MODES 4
/* Indexes into state array (0..3 head costs) */
#define QR_VER 4 /* Version */
#define QR_N_END 5 /* Numeric end index */
#define QR_N_COST 6 /* Numeric cost */
#define QR_A_END 7 /* Alpha end index */
#define QR_A_COST 8 /* Alpha cost */
#define QR_A_PCENT 9 /* Alpha 2nd char percent (GS1-specific) */
#define QR_A_PCCNT 10 /* Alpha total percent count (GS1-specific) */
2019-12-17 06:31:52 +13:00
/* Costs set to this for invalid MICROQR modes for versions M1 and M2.
* 128 is the max number of data bits for M4-L (ISO/IEC 18004:2015 Table 7) */
#define QR_MICROQR_MAX 774 /* (128 + 1) * QR_MULT */
/* Initial mode costs */
static unsigned int *qr_head_costs(unsigned int state[11]) {
static const unsigned int head_costs[7][QR_NUM_MODES] = {
2019-12-17 06:31:52 +13:00
/* N A B K */
{ (10 + 4) * QR_MULT, (9 + 4) * QR_MULT, (8 + 4) * QR_MULT, (8 + 4) * QR_MULT, }, /* QR */
{ (12 + 4) * QR_MULT, (11 + 4) * QR_MULT, (16 + 4) * QR_MULT, (10 + 4) * QR_MULT, },
{ (14 + 4) * QR_MULT, (13 + 4) * QR_MULT, (16 + 4) * QR_MULT, (12 + 4) * QR_MULT, },
{ 3 * QR_MULT, QR_MICROQR_MAX, QR_MICROQR_MAX, QR_MICROQR_MAX, }, /* MICROQR */
{ (4 + 1) * QR_MULT, (3 + 1) * QR_MULT, QR_MICROQR_MAX, QR_MICROQR_MAX, },
{ (5 + 2) * QR_MULT, (4 + 2) * QR_MULT, (4 + 2) * QR_MULT, (3 + 2) * QR_MULT, },
{ (6 + 3) * QR_MULT, (5 + 3) * QR_MULT, (5 + 3) * QR_MULT, (4 + 3) * QR_MULT, }
};
int version;
/* Head costs kept in states 0..3 */
version = state[QR_VER];
if (version < RMQR_VERSION) { /* QRCODE */
if (version < 10) {
memcpy(state, head_costs[0], QR_NUM_MODES * sizeof(unsigned int));
2019-12-17 06:31:52 +13:00
} else if (version < 27) {
memcpy(state, head_costs[1], QR_NUM_MODES * sizeof(unsigned int));
2019-12-17 06:31:52 +13:00
} else {
memcpy(state, head_costs[2], QR_NUM_MODES * sizeof(unsigned int));
}
2019-12-17 06:31:52 +13:00
} else if (version < MICROQR_VERSION) { /* RMQR */
version -= RMQR_VERSION;
state[QR_N] = (rmqr_numeric_cci[version] + 3) * QR_MULT;
state[QR_A] = (rmqr_alphanum_cci[version] + 3) * QR_MULT;
state[QR_B] = (rmqr_byte_cci[version] + 3) * QR_MULT;
state[QR_K] = (rmqr_kanji_cci[version] + 3) * QR_MULT;
} else { /* MICROQR */
memcpy(state, head_costs[3 + (version - MICROQR_VERSION)], QR_NUM_MODES * sizeof(unsigned int));
}
2016-02-17 23:52:48 +13:00
2019-12-17 06:31:52 +13:00
return state;
}
/* Calculate optimized encoding modes. Adapted from Project Nayuki */
static void qr_define_mode(char mode[], const unsigned int ddata[], const int length, const int gs1,
const int version, const int debug_print) {
/*
* Copyright (c) Project Nayuki. (MIT License)
* https://www.nayuki.io/page/qr-code-generator-library
*
* Permission is hereby granted, free of charge, to any person obtaining a copy of
* this software and associated documentation files (the "Software"), to deal in
* the Software without restriction, including without limitation the rights to
* use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of
* the Software, and to permit persons to whom the Software is furnished to do so,
* subject to the following conditions:
* - The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*/
unsigned int state[11] = {
0 /*N*/, 0 /*A*/, 0 /*B*/, 0 /*K*/, /* Head/switch costs */
0 /*version*/,
0 /*numeric_end*/, 0 /*numeric_cost*/, 0 /*alpha_end*/, 0 /*alpha_cost*/, 0 /*alpha_pcent*/, 0 /*alpha_pccnt*/
};
int m1, m2;
2019-12-17 06:31:52 +13:00
int i, j, k;
unsigned int min_cost;
char cur_mode;
unsigned int prev_costs[QR_NUM_MODES];
unsigned int cur_costs[QR_NUM_MODES];
char (*char_modes)[QR_NUM_MODES] = (char (*)[QR_NUM_MODES]) z_alloca(QR_NUM_MODES * length);
state[QR_VER] = (unsigned int) version;
/* char_modes[i][j] represents the mode to encode the code point at index i such that the final segment
ends in qr_mode_types[j] and the total number of bits is minimized over all possible choices */
memset(char_modes, 0, QR_NUM_MODES * length);
/* At the beginning of each iteration of the loop below, prev_costs[j] is the minimum number of 1/6 (1/QR_MULT)
* bits needed to encode the entire string prefix of length i, and end in qr_mode_types[j] */
memcpy(prev_costs, qr_head_costs(state), QR_NUM_MODES * sizeof(unsigned int));
#ifdef QR_DEBUG_DEFINE_MODE
printf(" head");
for (j = 0; j < QR_NUM_MODES; j++) {
printf(" %c(%c)=%d", qr_mode_types[j], char_modes[0][j], prev_costs[j]);
}
printf("\n");
#endif
/* Calculate costs using dynamic programming */
for (i = 0; i < length; i++) {
memset(cur_costs, 0, QR_NUM_MODES * sizeof(unsigned int));
m1 = version == MICROQR_VERSION;
m2 = version == MICROQR_VERSION + 1;
if (ddata[i] > 0xFF) {
cur_costs[QR_B] = prev_costs[QR_B] + ((m1 || m2) ? QR_MICROQR_MAX : 96); /* 16 * QR_MULT */
char_modes[i][QR_B] = 'B';
cur_costs[QR_K] = prev_costs[QR_K] + ((m1 || m2) ? QR_MICROQR_MAX : 78); /* 13 * QR_MULT */
char_modes[i][QR_K] = 'K';
} else {
if (qr_in_numeric(ddata, length, i, &state[QR_N_END], &state[QR_N_COST])) {
cur_costs[QR_N] = prev_costs[QR_N] + state[QR_N_COST];
char_modes[i][QR_N] = 'N';
}
if (qr_in_alpha(ddata, length, i, &state[QR_A_END], &state[QR_A_COST], &state[QR_A_PCENT],
&state[QR_A_PCCNT], gs1)) {
cur_costs[QR_A] = prev_costs[QR_A] + (m1 ? QR_MICROQR_MAX : state[QR_A_COST]);
char_modes[i][QR_A] = 'A';
}
cur_costs[QR_B] = prev_costs[QR_B] + ((m1 || m2) ? QR_MICROQR_MAX : 48); /* 8 * QR_MULT */
char_modes[i][QR_B] = 'B';
}
/* Start new segment at the end to switch modes */
for (j = 0; j < QR_NUM_MODES; j++) { /* To mode */
for (k = 0; k < QR_NUM_MODES; k++) { /* From mode */
if (j != k && char_modes[i][k]) {
const unsigned int new_cost = cur_costs[k] + state[j]; /* Switch costs same as head costs */
if (!char_modes[i][j] || new_cost < cur_costs[j]) {
cur_costs[j] = new_cost;
char_modes[i][j] = qr_mode_types[k];
}
}
}
2019-12-17 06:31:52 +13:00
}
#ifdef QR_DEBUG_DEFINE_MODE
{
int min_j = 0;
printf(" % 4d: curr", i);
for (j = 0; j < QR_NUM_MODES; j++) {
printf(" %c(%c)=%d", qr_mode_types[j], char_modes[i][j], cur_costs[j]);
if (cur_costs[j] < cur_costs[min_j]) min_j = j;
}
printf(" min %c(%c)=%d\n", qr_mode_types[min_j], char_modes[i][min_j], cur_costs[min_j]);
}
#endif
memcpy(prev_costs, cur_costs, QR_NUM_MODES * sizeof(unsigned int));
}
/* Find optimal ending mode */
min_cost = prev_costs[0];
cur_mode = qr_mode_types[0];
for (i = 1; i < QR_NUM_MODES; i++) {
if (prev_costs[i] < min_cost) {
min_cost = prev_costs[i];
cur_mode = qr_mode_types[i];
}
}
/* Get optimal mode for each code point by tracing backwards */
for (i = length - 1; i >= 0; i--) {
j = posn(qr_mode_types, cur_mode);
cur_mode = char_modes[i][j];
mode[i] = cur_mode;
}
if (debug_print) {
printf(" Mode: %.*s\n", length, mode);
}
2016-02-17 23:52:48 +13:00
}
/* Returns mode indicator based on version and mode */
static int qr_mode_indicator(const int version, const int mode) {
static const char mode_indicators[6][QR_NUM_MODES] = {
/*N A B K */
{ 1, 2, 4, 8, }, /* QRCODE */
{ 1, 2, 3, 4, }, /* RMQR */
{ 0, 0, 0, 0, }, /* MICROQR */
{ 0, 1, 0, 0, },
{ 0, 1, 2, 3, },
{ 0, 1, 2, 3, },
2019-12-17 06:31:52 +13:00
};
int mode_index = posn(qr_mode_types, (const char) mode);
if (version < RMQR_VERSION) {
return mode_indicators[0][mode_index]; /* QRCODE */
}
if (version < MICROQR_VERSION) {
return mode_indicators[1][mode_index] /* RMQR */;
}
return mode_indicators[2 + version - MICROQR_VERSION][mode_index]; /* MICROQR */
2019-12-17 06:31:52 +13:00
}
2019-12-17 06:31:52 +13:00
/* Return mode indicator bits based on version */
static int qr_mode_bits(const int version) {
2019-12-17 06:31:52 +13:00
if (version < RMQR_VERSION) {
return 4; /* QRCODE */
}
2019-12-17 06:31:52 +13:00
if (version < MICROQR_VERSION) {
return 3; /* RMQR */
}
return version - MICROQR_VERSION; /* MICROQR */
}
/* Return character count indicator bits based on version and mode */
static int qr_cci_bits(const int version, const int mode) {
static const char cci_bits[7][QR_NUM_MODES] = {
2019-12-17 06:31:52 +13:00
/* N A B K */
{ 10, 9, 8, 8, }, /* QRCODE */
{ 12, 11, 16, 10, },
{ 14, 13, 16, 12, },
{ 3, 0, 0, 0, }, /* MICROQR */
{ 4, 3, 0, 0, },
{ 5, 4, 4, 3, },
{ 6, 5, 5, 4, }
};
static const unsigned char *const rmqr_ccis[QR_NUM_MODES] = {
2019-12-17 06:31:52 +13:00
rmqr_numeric_cci, rmqr_alphanum_cci, rmqr_byte_cci, rmqr_kanji_cci,
};
int mode_index = posn(qr_mode_types, (const char) mode);
2019-12-17 06:31:52 +13:00
if (version < RMQR_VERSION) { /* QRCODE */
if (version < 10) {
return cci_bits[0][mode_index];
}
if (version < 27) {
return cci_bits[1][mode_index];
}
return cci_bits[2][mode_index];
}
if (version < MICROQR_VERSION) { /* RMQR */
return rmqr_ccis[mode_index][version - RMQR_VERSION];
}
return cci_bits[3 + (version - MICROQR_VERSION)][mode_index]; /* MICROQR */
}
/* Returns terminator bits based on version */
static int qr_terminator_bits(const int version) {
2019-12-17 06:31:52 +13:00
if (version < RMQR_VERSION) {
return 4; /* QRCODE */
}
if (version < MICROQR_VERSION) {
return 3; /* RMQR */
}
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return 3 + (version - MICROQR_VERSION) * 2; /* MICROQR */
2016-02-17 23:52:48 +13:00
}
/* Convert input data to a binary stream and add padding */
static int qr_binary(char binary[], int bp, const int version, const char mode[],
const unsigned int ddata[], const int length, const int gs1,
const int eci, const int debug_print) {
int position = 0;
int i;
int modebits;
int percent = 0;
int percent_count;
2016-02-17 23:52:48 +13:00
if (eci != 0) { /* Not applicable to MICROQR */
bp = bin_append_posn(7, version < RMQR_VERSION ? 4 : 3, binary, bp); /* ECI (Table 4) */
if (eci <= 127) {
bp = bin_append_posn(eci, 8, binary, bp); /* 000000 to 000127 */
} else if (eci <= 16383) {
bp = bin_append_posn(0x8000 + eci, 16, binary, bp); /* 000128 to 016383 */
} else {
bp = bin_append_posn(0xC00000 + eci, 24, binary, bp); /* 016384 to 999999 */
}
}
2016-02-17 23:52:48 +13:00
modebits = qr_mode_bits(version);
do {
2017-09-11 03:03:09 +12:00
char data_block = mode[position];
int short_data_block_length = 0;
int double_byte = 0;
do {
if (data_block == 'B' && ddata[position + short_data_block_length] > 0xFF) {
double_byte++;
}
short_data_block_length++;
} while (((short_data_block_length + position) < length)
&& (mode[position + short_data_block_length] == data_block));
/* Mode indicator */
if (modebits) {
bp = bin_append_posn(qr_mode_indicator(version, data_block), modebits, binary, bp);
}
switch (data_block) {
case 'K':
/* Kanji mode */
/* Character count indicator */
bp = bin_append_posn(short_data_block_length, qr_cci_bits(version, data_block), binary, bp);
if (debug_print) {
printf("Kanji block (length %d)\n\t", short_data_block_length);
}
/* Character representation */
for (i = 0; i < short_data_block_length; i++) {
unsigned int jis = ddata[position + i];
int prod;
2017-10-24 08:37:52 +13:00
if (jis >= 0x8140 && jis <= 0x9ffc)
jis -= 0x8140;
2017-10-24 08:37:52 +13:00
else if (jis >= 0xe040 && jis <= 0xebbf)
jis -= 0xc140;
2017-10-24 08:37:52 +13:00
prod = ((jis >> 8) * 0xc0) + (jis & 0xff);
2017-10-24 08:37:52 +13:00
bp = bin_append_posn(prod, 13, binary, bp);
if (debug_print) {
2019-12-17 06:31:52 +13:00
printf("0x%04X ", prod);
}
}
if (debug_print) {
EAN/UPC: add quiet zone indicators option (API `output_options` `EANUPC_GUARD_WHITESPACE`, CLI `--guardwhitespace`) (ticket #287) EAN-2/EAN-5: HRT now at top instead of at bottom for standalones, following BWIPP CLI: batch mode: don't close input if stdin EAN/UPC: fix excess 1X to right of add-ons Composites: fix excess whitespace; fix quiet zone calcs to allow for linear shifting CLI: use own (Wine) version of `CommandLineToArgvW()` to avoid loading "shell32.dll" Move "font.h" -> "raster_font.h" EPS/SVG: use new `out_putsf()` func to output floats, avoiding trailing zeroes & locale dependency EPS: simplify "TR" formula SVG: change font from "Helvetica, sans serif" to "OCR-B, monospace"; use single "<path>" instead of multiple "<rect>"s to draw boxes (reduces file size) Add `EMBED_VECTOR_FONT` to `output_options` (CLI `--embedfont`) to enable embedding of font in vector output - currently only for SVG output of EAN/UPC GUI: use "OCR-B" font for EAN/UPC and "Arimo" for all others (was "Helvetica" for both); paint background of screen preview light grey so as contrasts with whitespace and quiet zones EMF: prefix funcs with `emf_`; simplify string `halign` handling large: rename `large_int` -> `large_uint` CODE128/common: move `c128_hrt_cpy_iso8859_1()` to `hrt_cpy_iso8859_1()` and add `ZINT_WARN_HRT_TRUNCATED` warning (for future use) Various symbologies: replace `printf()` with `fputs()` (symbol->debug) QRCODE: better assert(), removing a NOLINT (2 left) CLI: add some more barcode synonyms for DBAR common: various fiddlings CMake: don't include png.c unless ZINT_USE_PNG (avoids clang warning)
2023-06-12 12:25:55 +12:00
fputc('\n', stdout);
}
break;
case 'B':
/* Byte mode */
/* Character count indicator */
bp = bin_append_posn(short_data_block_length + double_byte, qr_cci_bits(version, data_block), binary,
bp);
if (debug_print) {
printf("Byte block (length %d)\n\t", short_data_block_length + double_byte);
}
/* Character representation */
for (i = 0; i < short_data_block_length; i++) {
unsigned int byte = ddata[position + i];
2016-02-17 23:52:48 +13:00
bp = bin_append_posn(byte, byte > 0xFF ? 16 : 8, binary, bp);
if (debug_print) {
printf("0x%02X(%d) ", byte, (int) byte);
}
}
if (debug_print) {
EAN/UPC: add quiet zone indicators option (API `output_options` `EANUPC_GUARD_WHITESPACE`, CLI `--guardwhitespace`) (ticket #287) EAN-2/EAN-5: HRT now at top instead of at bottom for standalones, following BWIPP CLI: batch mode: don't close input if stdin EAN/UPC: fix excess 1X to right of add-ons Composites: fix excess whitespace; fix quiet zone calcs to allow for linear shifting CLI: use own (Wine) version of `CommandLineToArgvW()` to avoid loading "shell32.dll" Move "font.h" -> "raster_font.h" EPS/SVG: use new `out_putsf()` func to output floats, avoiding trailing zeroes & locale dependency EPS: simplify "TR" formula SVG: change font from "Helvetica, sans serif" to "OCR-B, monospace"; use single "<path>" instead of multiple "<rect>"s to draw boxes (reduces file size) Add `EMBED_VECTOR_FONT` to `output_options` (CLI `--embedfont`) to enable embedding of font in vector output - currently only for SVG output of EAN/UPC GUI: use "OCR-B" font for EAN/UPC and "Arimo" for all others (was "Helvetica" for both); paint background of screen preview light grey so as contrasts with whitespace and quiet zones EMF: prefix funcs with `emf_`; simplify string `halign` handling large: rename `large_int` -> `large_uint` CODE128/common: move `c128_hrt_cpy_iso8859_1()` to `hrt_cpy_iso8859_1()` and add `ZINT_WARN_HRT_TRUNCATED` warning (for future use) Various symbologies: replace `printf()` with `fputs()` (symbol->debug) QRCODE: better assert(), removing a NOLINT (2 left) CLI: add some more barcode synonyms for DBAR common: various fiddlings CMake: don't include png.c unless ZINT_USE_PNG (avoids clang warning)
2023-06-12 12:25:55 +12:00
fputc('\n', stdout);
}
break;
case 'A':
/* Alphanumeric mode */
percent_count = 0;
2019-12-17 06:31:52 +13:00
if (gs1) {
for (i = 0; i < short_data_block_length; i++) {
if (ddata[position + i] == '%') {
2019-12-17 06:31:52 +13:00
percent_count++;
}
}
}
/* Character count indicator */
bp = bin_append_posn(short_data_block_length + percent_count, qr_cci_bits(version, data_block),
binary, bp);
if (debug_print) {
printf("Alpha block (length %d)\n\t", short_data_block_length + percent_count);
}
/* Character representation */
i = 0;
while (i < short_data_block_length) {
int count;
int first = 0, second = 0, prod;
if (percent == 0) {
if (gs1 && (ddata[position + i] == '%')) {
first = QR_PERCENT;
second = QR_PERCENT;
count = 2;
prod = (first * 45) + second;
i++;
} else {
if (gs1 && ddata[position + i] == '\x1D') {
first = QR_PERCENT; /* FNC1 */
} else {
first = qr_alphanumeric[ddata[position + i] - 32];
}
count = 1;
i++;
prod = first;
if (i < short_data_block_length && mode[position + i] == 'A') {
if (gs1 && (ddata[position + i] == '%')) {
second = QR_PERCENT;
count = 2;
prod = (first * 45) + second;
percent = 1;
} else {
if (gs1 && ddata[position + i] == '\x1D') {
second = QR_PERCENT; /* FNC1 */
} else {
second = qr_alphanumeric[ddata[position + i] - 32];
}
count = 2;
i++;
prod = (first * 45) + second;
}
}
}
} else {
first = QR_PERCENT;
count = 1;
i++;
prod = first;
percent = 0;
if (i < short_data_block_length && mode[position + i] == 'A') {
if (gs1 && (ddata[position + i] == '%')) {
second = QR_PERCENT;
count = 2;
prod = (first * 45) + second;
percent = 1;
} else {
if (gs1 && ddata[position + i] == '\x1D') {
second = QR_PERCENT; /* FNC1 */
} else {
second = qr_alphanumeric[ddata[position + i] - 32];
}
count = 2;
i++;
prod = (first * 45) + second;
}
}
}
bp = bin_append_posn(prod, 1 + (5 * count), binary, bp);
if (debug_print) {
2019-12-17 06:31:52 +13:00
printf("0x%X ", prod);
}
2019-12-17 06:31:52 +13:00
}
if (debug_print) {
EAN/UPC: add quiet zone indicators option (API `output_options` `EANUPC_GUARD_WHITESPACE`, CLI `--guardwhitespace`) (ticket #287) EAN-2/EAN-5: HRT now at top instead of at bottom for standalones, following BWIPP CLI: batch mode: don't close input if stdin EAN/UPC: fix excess 1X to right of add-ons Composites: fix excess whitespace; fix quiet zone calcs to allow for linear shifting CLI: use own (Wine) version of `CommandLineToArgvW()` to avoid loading "shell32.dll" Move "font.h" -> "raster_font.h" EPS/SVG: use new `out_putsf()` func to output floats, avoiding trailing zeroes & locale dependency EPS: simplify "TR" formula SVG: change font from "Helvetica, sans serif" to "OCR-B, monospace"; use single "<path>" instead of multiple "<rect>"s to draw boxes (reduces file size) Add `EMBED_VECTOR_FONT` to `output_options` (CLI `--embedfont`) to enable embedding of font in vector output - currently only for SVG output of EAN/UPC GUI: use "OCR-B" font for EAN/UPC and "Arimo" for all others (was "Helvetica" for both); paint background of screen preview light grey so as contrasts with whitespace and quiet zones EMF: prefix funcs with `emf_`; simplify string `halign` handling large: rename `large_int` -> `large_uint` CODE128/common: move `c128_hrt_cpy_iso8859_1()` to `hrt_cpy_iso8859_1()` and add `ZINT_WARN_HRT_TRUNCATED` warning (for future use) Various symbologies: replace `printf()` with `fputs()` (symbol->debug) QRCODE: better assert(), removing a NOLINT (2 left) CLI: add some more barcode synonyms for DBAR common: various fiddlings CMake: don't include png.c unless ZINT_USE_PNG (avoids clang warning)
2023-06-12 12:25:55 +12:00
fputc('\n', stdout);
}
break;
case 'N':
/* Numeric mode */
/* Character count indicator */
bp = bin_append_posn(short_data_block_length, qr_cci_bits(version, data_block), binary, bp);
if (debug_print) {
printf("Number block (length %d)\n\t", short_data_block_length);
}
/* Character representation */
i = 0;
while (i < short_data_block_length) {
int count;
2017-09-11 03:03:09 +12:00
int first = 0, prod;
first = ctoi((const char) ddata[position + i]);
count = 1;
prod = first;
if (i + 1 < short_data_block_length && mode[position + i + 1] == 'N') {
int second = ctoi((const char) ddata[position + i + 1]);
count = 2;
prod = (prod * 10) + second;
if (i + 2 < short_data_block_length && mode[position + i + 2] == 'N') {
int third = ctoi((const char) ddata[position + i + 2]);
count = 3;
prod = (prod * 10) + third;
}
}
bp = bin_append_posn(prod, 1 + (3 * count), binary, bp);
if (debug_print) {
2019-12-17 06:31:52 +13:00
printf("0x%X(%d) ", prod, prod);
}
i += count;
};
if (debug_print) {
EAN/UPC: add quiet zone indicators option (API `output_options` `EANUPC_GUARD_WHITESPACE`, CLI `--guardwhitespace`) (ticket #287) EAN-2/EAN-5: HRT now at top instead of at bottom for standalones, following BWIPP CLI: batch mode: don't close input if stdin EAN/UPC: fix excess 1X to right of add-ons Composites: fix excess whitespace; fix quiet zone calcs to allow for linear shifting CLI: use own (Wine) version of `CommandLineToArgvW()` to avoid loading "shell32.dll" Move "font.h" -> "raster_font.h" EPS/SVG: use new `out_putsf()` func to output floats, avoiding trailing zeroes & locale dependency EPS: simplify "TR" formula SVG: change font from "Helvetica, sans serif" to "OCR-B, monospace"; use single "<path>" instead of multiple "<rect>"s to draw boxes (reduces file size) Add `EMBED_VECTOR_FONT` to `output_options` (CLI `--embedfont`) to enable embedding of font in vector output - currently only for SVG output of EAN/UPC GUI: use "OCR-B" font for EAN/UPC and "Arimo" for all others (was "Helvetica" for both); paint background of screen preview light grey so as contrasts with whitespace and quiet zones EMF: prefix funcs with `emf_`; simplify string `halign` handling large: rename `large_int` -> `large_uint` CODE128/common: move `c128_hrt_cpy_iso8859_1()` to `hrt_cpy_iso8859_1()` and add `ZINT_WARN_HRT_TRUNCATED` warning (for future use) Various symbologies: replace `printf()` with `fputs()` (symbol->debug) QRCODE: better assert(), removing a NOLINT (2 left) CLI: add some more barcode synonyms for DBAR common: various fiddlings CMake: don't include png.c unless ZINT_USE_PNG (avoids clang warning)
2023-06-12 12:25:55 +12:00
fputc('\n', stdout);
}
break;
}
2016-02-17 23:52:48 +13:00
position += short_data_block_length;
} while (position < length);
2016-02-17 23:52:48 +13:00
return bp;
}
/* Call `qr_binary()` for each segment, dealing with Structured Append and GS1 beforehand and padding afterwards */
static int qr_binary_segs(unsigned char datastream[], const int version, const int target_codewords,
const char mode[], const unsigned int ddata[], const struct zint_seg segs[], const int seg_count,
const struct zint_structapp *p_structapp, const int gs1, const int est_binlen, const int debug_print) {
int i, j;
const unsigned int *dd = ddata;
const char *m = mode;
int bp = 0;
int termbits, padbits;
int current_bytes;
int toggle;
char *binary = (char *) z_alloca(est_binlen + 12);
*binary = '\0';
if (p_structapp) {
bp = bin_append_posn(3, 4, binary, bp); /* Structured Append indicator */
bp = bin_append_posn(p_structapp->index - 1, 4, binary, bp);
bp = bin_append_posn(p_structapp->count - 1, 4, binary, bp);
bp = bin_append_posn(to_int((const unsigned char *) p_structapp->id, (int) strlen(p_structapp->id)), 8,
binary, bp); /* Parity */
}
if (gs1) { /* Not applicable to MICROQR */
if (version < RMQR_VERSION) {
bp = bin_append_posn(5, 4, binary, bp); /* FNC1 */
} else {
bp = bin_append_posn(5, 3, binary, bp);
}
}
for (i = 0; i < seg_count; i++) {
bp = qr_binary(binary, bp, version, m, dd, segs[i].length, gs1, segs[i].eci, debug_print);
m += segs[i].length;
dd += segs[i].length;
}
2019-12-17 06:31:52 +13:00
if (version >= MICROQR_VERSION && version < MICROQR_VERSION + 4) {
/* MICROQR does its own terminating/padding */
Performance improvements for linear encoding and raster output - use fixed-length string tables (mostly) instead of (char *) pointer ones (saves ~40K) - re-use C128Table for CODABLOCKF and CODE16K (required removal of Stop character and extra CODE16K-only entry) - use pointer to destination and copy (memcpy/strcpy(), bin_append_posn()) instead of concatenating (strcat()) (mostly) - replace last remaining bin_append()s with bin_append_posn(); bin_append() removed - add length arg to toupper() and expand() (avoids strlen()) - change is_sane() to use table-based flags (avoids an iteration) - rename lookup() to is_sane_lookup() and change to check and return posns and use in pointer to destination loops (avoids strcat()s) - remove special case PHARMA in expand() (dealt with in pharma()) - make #define SILVER/CALCIUM/TECHNETIUM/KRSET etc static strings - replace strchr() -> posn() - CODE128: populate destination once in checksum loop; re-use and export some more routines (c128_set_a/b/c(), c128_put_in_set()) for sharing; prefix defines (SHIFTA -> C128_SHIFTA etc) and existing exported routines - use factor XOR toggle trick in checksum calcs (avoids branch) - raster.c: fill out single 1-pixel row and copy using new draw_bar_line(), copy_bar_line() routines; similarly in buffer_plot compare previous line & copy if same (same technique as used to improve non-half-integer scaling, significant performance increase, (c) codemonkey82); also done for PNG (BMP/GIF/PCX/TIFF not done) - raster/vector/output.c: shorten "output_" prefix -> "out_"; sync vector to other raster changes to try to keep source files similar - 2of5.c: prefix "c25_" JAPANPOST: return error if input data truncated (backward incompatible) DAFT: max chars 50 -> 100 common.c: istwodigit() -> is_twodigit() common.c/emf.c/output.c: use some further stripf()s (MSVC6 float variations) library.c: new check_output_args() helper zint.h: add BARCODE_LAST marker and use in library.c QRCODE: remove a NOLINT (requires clang-tidy-13), one remaining CMake: separate no-optimize from ZINT_DEBUG into new ZINT_NOOPT option
2021-10-21 11:05:30 +13:00
memcpy(datastream, binary, bp);
return bp;
}
2016-02-17 23:52:48 +13:00
2019-12-17 06:31:52 +13:00
/* Terminator */
termbits = 8 - bp % 8;
2019-12-17 06:31:52 +13:00
if (termbits == 8) {
termbits = 0;
}
current_bytes = (bp + termbits) / 8;
2019-12-17 06:31:52 +13:00
if (termbits || current_bytes < target_codewords) {
int max_termbits = qr_terminator_bits(version);
2019-12-17 06:31:52 +13:00
termbits = termbits < max_termbits && current_bytes == target_codewords ? termbits : max_termbits;
bp = bin_append_posn(0, termbits, binary, bp);
}
2016-02-17 23:52:48 +13:00
/* Padding bits */
padbits = 8 - bp % 8;
2019-12-17 06:31:52 +13:00
if (padbits == 8) {
padbits = 0;
}
if (padbits) {
current_bytes = (bp + padbits) / 8;
(void) bin_append_posn(0, padbits, binary, bp); /* Last use so not setting bp */
}
2016-02-17 23:52:48 +13:00
if (debug_print) printf("Terminated binary (%d): %.*s (padbits %d)\n", bp, bp, binary, padbits);
/* Put data into 8-bit codewords */
for (i = 0; i < current_bytes; i++) {
int p;
j = i * 8;
datastream[i] = 0x00;
for (p = 0; p < 8; p++) {
if (binary[j + p] == '1') {
datastream[i] |= (0x80 >> p);
}
}
}
/* Add pad codewords */
toggle = 0;
for (i = current_bytes; i < target_codewords; i++) {
if (toggle == 0) {
datastream[i] = 0xec;
toggle = 1;
} else {
datastream[i] = 0x11;
toggle = 0;
}
}
2016-02-17 23:52:48 +13:00
if (debug_print) {
printf("Resulting codewords (%d):\n\t", target_codewords);
for (i = 0; i < target_codewords; i++) {
2019-12-17 06:31:52 +13:00
printf("0x%02X ", datastream[i]);
}
EAN/UPC: add quiet zone indicators option (API `output_options` `EANUPC_GUARD_WHITESPACE`, CLI `--guardwhitespace`) (ticket #287) EAN-2/EAN-5: HRT now at top instead of at bottom for standalones, following BWIPP CLI: batch mode: don't close input if stdin EAN/UPC: fix excess 1X to right of add-ons Composites: fix excess whitespace; fix quiet zone calcs to allow for linear shifting CLI: use own (Wine) version of `CommandLineToArgvW()` to avoid loading "shell32.dll" Move "font.h" -> "raster_font.h" EPS/SVG: use new `out_putsf()` func to output floats, avoiding trailing zeroes & locale dependency EPS: simplify "TR" formula SVG: change font from "Helvetica, sans serif" to "OCR-B, monospace"; use single "<path>" instead of multiple "<rect>"s to draw boxes (reduces file size) Add `EMBED_VECTOR_FONT` to `output_options` (CLI `--embedfont`) to enable embedding of font in vector output - currently only for SVG output of EAN/UPC GUI: use "OCR-B" font for EAN/UPC and "Arimo" for all others (was "Helvetica" for both); paint background of screen preview light grey so as contrasts with whitespace and quiet zones EMF: prefix funcs with `emf_`; simplify string `halign` handling large: rename `large_int` -> `large_uint` CODE128/common: move `c128_hrt_cpy_iso8859_1()` to `hrt_cpy_iso8859_1()` and add `ZINT_WARN_HRT_TRUNCATED` warning (for future use) Various symbologies: replace `printf()` with `fputs()` (symbol->debug) QRCODE: better assert(), removing a NOLINT (2 left) CLI: add some more barcode synonyms for DBAR common: various fiddlings CMake: don't include png.c unless ZINT_USE_PNG (avoids clang warning)
2023-06-12 12:25:55 +12:00
fputc('\n', stdout);
}
Performance improvements for linear encoding and raster output - use fixed-length string tables (mostly) instead of (char *) pointer ones (saves ~40K) - re-use C128Table for CODABLOCKF and CODE16K (required removal of Stop character and extra CODE16K-only entry) - use pointer to destination and copy (memcpy/strcpy(), bin_append_posn()) instead of concatenating (strcat()) (mostly) - replace last remaining bin_append()s with bin_append_posn(); bin_append() removed - add length arg to toupper() and expand() (avoids strlen()) - change is_sane() to use table-based flags (avoids an iteration) - rename lookup() to is_sane_lookup() and change to check and return posns and use in pointer to destination loops (avoids strcat()s) - remove special case PHARMA in expand() (dealt with in pharma()) - make #define SILVER/CALCIUM/TECHNETIUM/KRSET etc static strings - replace strchr() -> posn() - CODE128: populate destination once in checksum loop; re-use and export some more routines (c128_set_a/b/c(), c128_put_in_set()) for sharing; prefix defines (SHIFTA -> C128_SHIFTA etc) and existing exported routines - use factor XOR toggle trick in checksum calcs (avoids branch) - raster.c: fill out single 1-pixel row and copy using new draw_bar_line(), copy_bar_line() routines; similarly in buffer_plot compare previous line & copy if same (same technique as used to improve non-half-integer scaling, significant performance increase, (c) codemonkey82); also done for PNG (BMP/GIF/PCX/TIFF not done) - raster/vector/output.c: shorten "output_" prefix -> "out_"; sync vector to other raster changes to try to keep source files similar - 2of5.c: prefix "c25_" JAPANPOST: return error if input data truncated (backward incompatible) DAFT: max chars 50 -> 100 common.c: istwodigit() -> is_twodigit() common.c/emf.c/output.c: use some further stripf()s (MSVC6 float variations) library.c: new check_output_args() helper zint.h: add BARCODE_LAST marker and use in library.c QRCODE: remove a NOLINT (requires clang-tidy-13), one remaining CMake: separate no-optimize from ZINT_DEBUG into new ZINT_NOOPT option
2021-10-21 11:05:30 +13:00
return 0; /* Not used */
2016-02-17 23:52:48 +13:00
}
/* Split data into blocks, add error correction and then interleave the blocks and error correction data */
static void qr_add_ecc(unsigned char fullstream[], const unsigned char datastream[], const int version,
const int data_cw, const int blocks, const int debug_print) {
int ecc_cw;
int short_data_block_length;
int qty_long_blocks;
int qty_short_blocks;
int ecc_block_length;
int i, j, length_this_block, in_posn;
rs_t rs;
unsigned char *data_block;
unsigned char *ecc_block;
unsigned char *interleaved_data;
unsigned char *interleaved_ecc;
if (version < RMQR_VERSION) {
ecc_cw = qr_total_codewords[version - 1] - data_cw;
} else {
ecc_cw = rmqr_total_codewords[version - RMQR_VERSION] - data_cw;
}
/* Suppress some clang-tidy clang-analyzer-core.UndefinedBinaryOperatorResult/uninitialized.Assign warnings */
assert(blocks > 0);
EAN/UPC: add quiet zone indicators option (API `output_options` `EANUPC_GUARD_WHITESPACE`, CLI `--guardwhitespace`) (ticket #287) EAN-2/EAN-5: HRT now at top instead of at bottom for standalones, following BWIPP CLI: batch mode: don't close input if stdin EAN/UPC: fix excess 1X to right of add-ons Composites: fix excess whitespace; fix quiet zone calcs to allow for linear shifting CLI: use own (Wine) version of `CommandLineToArgvW()` to avoid loading "shell32.dll" Move "font.h" -> "raster_font.h" EPS/SVG: use new `out_putsf()` func to output floats, avoiding trailing zeroes & locale dependency EPS: simplify "TR" formula SVG: change font from "Helvetica, sans serif" to "OCR-B, monospace"; use single "<path>" instead of multiple "<rect>"s to draw boxes (reduces file size) Add `EMBED_VECTOR_FONT` to `output_options` (CLI `--embedfont`) to enable embedding of font in vector output - currently only for SVG output of EAN/UPC GUI: use "OCR-B" font for EAN/UPC and "Arimo" for all others (was "Helvetica" for both); paint background of screen preview light grey so as contrasts with whitespace and quiet zones EMF: prefix funcs with `emf_`; simplify string `halign` handling large: rename `large_int` -> `large_uint` CODE128/common: move `c128_hrt_cpy_iso8859_1()` to `hrt_cpy_iso8859_1()` and add `ZINT_WARN_HRT_TRUNCATED` warning (for future use) Various symbologies: replace `printf()` with `fputs()` (symbol->debug) QRCODE: better assert(), removing a NOLINT (2 left) CLI: add some more barcode synonyms for DBAR common: various fiddlings CMake: don't include png.c unless ZINT_USE_PNG (avoids clang warning)
2023-06-12 12:25:55 +12:00
short_data_block_length = data_cw / blocks;
qty_long_blocks = data_cw % blocks;
qty_short_blocks = blocks - qty_long_blocks;
ecc_block_length = ecc_cw / blocks;
/* Suppress some clang-tidy clang-analyzer-core.UndefinedBinaryOperatorResult/uninitialized.Assign warnings */
Performance improvements for linear encoding and raster output - use fixed-length string tables (mostly) instead of (char *) pointer ones (saves ~40K) - re-use C128Table for CODABLOCKF and CODE16K (required removal of Stop character and extra CODE16K-only entry) - use pointer to destination and copy (memcpy/strcpy(), bin_append_posn()) instead of concatenating (strcat()) (mostly) - replace last remaining bin_append()s with bin_append_posn(); bin_append() removed - add length arg to toupper() and expand() (avoids strlen()) - change is_sane() to use table-based flags (avoids an iteration) - rename lookup() to is_sane_lookup() and change to check and return posns and use in pointer to destination loops (avoids strcat()s) - remove special case PHARMA in expand() (dealt with in pharma()) - make #define SILVER/CALCIUM/TECHNETIUM/KRSET etc static strings - replace strchr() -> posn() - CODE128: populate destination once in checksum loop; re-use and export some more routines (c128_set_a/b/c(), c128_put_in_set()) for sharing; prefix defines (SHIFTA -> C128_SHIFTA etc) and existing exported routines - use factor XOR toggle trick in checksum calcs (avoids branch) - raster.c: fill out single 1-pixel row and copy using new draw_bar_line(), copy_bar_line() routines; similarly in buffer_plot compare previous line & copy if same (same technique as used to improve non-half-integer scaling, significant performance increase, (c) codemonkey82); also done for PNG (BMP/GIF/PCX/TIFF not done) - raster/vector/output.c: shorten "output_" prefix -> "out_"; sync vector to other raster changes to try to keep source files similar - 2of5.c: prefix "c25_" JAPANPOST: return error if input data truncated (backward incompatible) DAFT: max chars 50 -> 100 common.c: istwodigit() -> is_twodigit() common.c/emf.c/output.c: use some further stripf()s (MSVC6 float variations) library.c: new check_output_args() helper zint.h: add BARCODE_LAST marker and use in library.c QRCODE: remove a NOLINT (requires clang-tidy-13), one remaining CMake: separate no-optimize from ZINT_DEBUG into new ZINT_NOOPT option
2021-10-21 11:05:30 +13:00
assert(short_data_block_length > 0);
EAN/UPC: add quiet zone indicators option (API `output_options` `EANUPC_GUARD_WHITESPACE`, CLI `--guardwhitespace`) (ticket #287) EAN-2/EAN-5: HRT now at top instead of at bottom for standalones, following BWIPP CLI: batch mode: don't close input if stdin EAN/UPC: fix excess 1X to right of add-ons Composites: fix excess whitespace; fix quiet zone calcs to allow for linear shifting CLI: use own (Wine) version of `CommandLineToArgvW()` to avoid loading "shell32.dll" Move "font.h" -> "raster_font.h" EPS/SVG: use new `out_putsf()` func to output floats, avoiding trailing zeroes & locale dependency EPS: simplify "TR" formula SVG: change font from "Helvetica, sans serif" to "OCR-B, monospace"; use single "<path>" instead of multiple "<rect>"s to draw boxes (reduces file size) Add `EMBED_VECTOR_FONT` to `output_options` (CLI `--embedfont`) to enable embedding of font in vector output - currently only for SVG output of EAN/UPC GUI: use "OCR-B" font for EAN/UPC and "Arimo" for all others (was "Helvetica" for both); paint background of screen preview light grey so as contrasts with whitespace and quiet zones EMF: prefix funcs with `emf_`; simplify string `halign` handling large: rename `large_int` -> `large_uint` CODE128/common: move `c128_hrt_cpy_iso8859_1()` to `hrt_cpy_iso8859_1()` and add `ZINT_WARN_HRT_TRUNCATED` warning (for future use) Various symbologies: replace `printf()` with `fputs()` (symbol->debug) QRCODE: better assert(), removing a NOLINT (2 left) CLI: add some more barcode synonyms for DBAR common: various fiddlings CMake: don't include png.c unless ZINT_USE_PNG (avoids clang warning)
2023-06-12 12:25:55 +12:00
assert(qty_long_blocks || qty_short_blocks);
data_block = (unsigned char *) z_alloca(short_data_block_length + 1);
ecc_block = (unsigned char *) z_alloca(ecc_block_length);
interleaved_data = (unsigned char *) z_alloca(data_cw);
interleaved_ecc = (unsigned char *) z_alloca(ecc_cw);
2016-02-17 23:52:48 +13:00
rs_init_gf(&rs, 0x11d);
rs_init_code(&rs, ecc_block_length, 0);
in_posn = 0;
2016-02-17 23:52:48 +13:00
for (i = 0; i < blocks; i++) {
if (i < qty_short_blocks) {
length_this_block = short_data_block_length;
} else {
length_this_block = short_data_block_length + 1;
}
for (j = 0; j < length_this_block; j++) {
/* This false-positive popped up with clang-tidy 14.0.1 */
data_block[j] = datastream[in_posn + j]; /* NOLINT(clang-analyzer-core.uninitialized.Assign) */
}
2016-02-17 23:52:48 +13:00
rs_encode(&rs, length_this_block, data_block, ecc_block);
2016-02-17 23:52:48 +13:00
if (debug_print) {
printf("Block %d: ", i + 1);
for (j = 0; j < length_this_block; j++) {
printf("%2X ", data_block[j]);
}
if (i < qty_short_blocks) {
EAN/UPC: add quiet zone indicators option (API `output_options` `EANUPC_GUARD_WHITESPACE`, CLI `--guardwhitespace`) (ticket #287) EAN-2/EAN-5: HRT now at top instead of at bottom for standalones, following BWIPP CLI: batch mode: don't close input if stdin EAN/UPC: fix excess 1X to right of add-ons Composites: fix excess whitespace; fix quiet zone calcs to allow for linear shifting CLI: use own (Wine) version of `CommandLineToArgvW()` to avoid loading "shell32.dll" Move "font.h" -> "raster_font.h" EPS/SVG: use new `out_putsf()` func to output floats, avoiding trailing zeroes & locale dependency EPS: simplify "TR" formula SVG: change font from "Helvetica, sans serif" to "OCR-B, monospace"; use single "<path>" instead of multiple "<rect>"s to draw boxes (reduces file size) Add `EMBED_VECTOR_FONT` to `output_options` (CLI `--embedfont`) to enable embedding of font in vector output - currently only for SVG output of EAN/UPC GUI: use "OCR-B" font for EAN/UPC and "Arimo" for all others (was "Helvetica" for both); paint background of screen preview light grey so as contrasts with whitespace and quiet zones EMF: prefix funcs with `emf_`; simplify string `halign` handling large: rename `large_int` -> `large_uint` CODE128/common: move `c128_hrt_cpy_iso8859_1()` to `hrt_cpy_iso8859_1()` and add `ZINT_WARN_HRT_TRUNCATED` warning (for future use) Various symbologies: replace `printf()` with `fputs()` (symbol->debug) QRCODE: better assert(), removing a NOLINT (2 left) CLI: add some more barcode synonyms for DBAR common: various fiddlings CMake: don't include png.c unless ZINT_USE_PNG (avoids clang warning)
2023-06-12 12:25:55 +12:00
fputs(" ", stdout);
}
EAN/UPC: add quiet zone indicators option (API `output_options` `EANUPC_GUARD_WHITESPACE`, CLI `--guardwhitespace`) (ticket #287) EAN-2/EAN-5: HRT now at top instead of at bottom for standalones, following BWIPP CLI: batch mode: don't close input if stdin EAN/UPC: fix excess 1X to right of add-ons Composites: fix excess whitespace; fix quiet zone calcs to allow for linear shifting CLI: use own (Wine) version of `CommandLineToArgvW()` to avoid loading "shell32.dll" Move "font.h" -> "raster_font.h" EPS/SVG: use new `out_putsf()` func to output floats, avoiding trailing zeroes & locale dependency EPS: simplify "TR" formula SVG: change font from "Helvetica, sans serif" to "OCR-B, monospace"; use single "<path>" instead of multiple "<rect>"s to draw boxes (reduces file size) Add `EMBED_VECTOR_FONT` to `output_options` (CLI `--embedfont`) to enable embedding of font in vector output - currently only for SVG output of EAN/UPC GUI: use "OCR-B" font for EAN/UPC and "Arimo" for all others (was "Helvetica" for both); paint background of screen preview light grey so as contrasts with whitespace and quiet zones EMF: prefix funcs with `emf_`; simplify string `halign` handling large: rename `large_int` -> `large_uint` CODE128/common: move `c128_hrt_cpy_iso8859_1()` to `hrt_cpy_iso8859_1()` and add `ZINT_WARN_HRT_TRUNCATED` warning (for future use) Various symbologies: replace `printf()` with `fputs()` (symbol->debug) QRCODE: better assert(), removing a NOLINT (2 left) CLI: add some more barcode synonyms for DBAR common: various fiddlings CMake: don't include png.c unless ZINT_USE_PNG (avoids clang warning)
2023-06-12 12:25:55 +12:00
fputs(" // ", stdout);
for (j = 0; j < ecc_block_length; j++) {
printf("%2X ", ecc_block[ecc_block_length - j - 1]);
}
EAN/UPC: add quiet zone indicators option (API `output_options` `EANUPC_GUARD_WHITESPACE`, CLI `--guardwhitespace`) (ticket #287) EAN-2/EAN-5: HRT now at top instead of at bottom for standalones, following BWIPP CLI: batch mode: don't close input if stdin EAN/UPC: fix excess 1X to right of add-ons Composites: fix excess whitespace; fix quiet zone calcs to allow for linear shifting CLI: use own (Wine) version of `CommandLineToArgvW()` to avoid loading "shell32.dll" Move "font.h" -> "raster_font.h" EPS/SVG: use new `out_putsf()` func to output floats, avoiding trailing zeroes & locale dependency EPS: simplify "TR" formula SVG: change font from "Helvetica, sans serif" to "OCR-B, monospace"; use single "<path>" instead of multiple "<rect>"s to draw boxes (reduces file size) Add `EMBED_VECTOR_FONT` to `output_options` (CLI `--embedfont`) to enable embedding of font in vector output - currently only for SVG output of EAN/UPC GUI: use "OCR-B" font for EAN/UPC and "Arimo" for all others (was "Helvetica" for both); paint background of screen preview light grey so as contrasts with whitespace and quiet zones EMF: prefix funcs with `emf_`; simplify string `halign` handling large: rename `large_int` -> `large_uint` CODE128/common: move `c128_hrt_cpy_iso8859_1()` to `hrt_cpy_iso8859_1()` and add `ZINT_WARN_HRT_TRUNCATED` warning (for future use) Various symbologies: replace `printf()` with `fputs()` (symbol->debug) QRCODE: better assert(), removing a NOLINT (2 left) CLI: add some more barcode synonyms for DBAR common: various fiddlings CMake: don't include png.c unless ZINT_USE_PNG (avoids clang warning)
2023-06-12 12:25:55 +12:00
fputc('\n', stdout);
}
2016-02-17 23:52:48 +13:00
for (j = 0; j < short_data_block_length; j++) {
/* And another with clang-tidy 14.0.6 */
interleaved_data[(j * blocks) + i] = data_block[j]; /* NOLINT(clang-analyzer-core.uninitialized.Assign) */
}
2016-02-17 23:52:48 +13:00
if (i >= qty_short_blocks) {
interleaved_data[(short_data_block_length * blocks) + (i - qty_short_blocks)]
= data_block[short_data_block_length];
}
2016-02-17 23:52:48 +13:00
for (j = 0; j < ecc_block_length; j++) {
interleaved_ecc[(j * blocks) + i] = ecc_block[ecc_block_length - j - 1];
}
2016-02-17 23:52:48 +13:00
in_posn += length_this_block;
}
2016-02-17 23:52:48 +13:00
for (j = 0; j < data_cw; j++) {
fullstream[j] = interleaved_data[j];
}
for (j = 0; j < ecc_cw; j++) {
EAN/UPC: add quiet zone indicators option (API `output_options` `EANUPC_GUARD_WHITESPACE`, CLI `--guardwhitespace`) (ticket #287) EAN-2/EAN-5: HRT now at top instead of at bottom for standalones, following BWIPP CLI: batch mode: don't close input if stdin EAN/UPC: fix excess 1X to right of add-ons Composites: fix excess whitespace; fix quiet zone calcs to allow for linear shifting CLI: use own (Wine) version of `CommandLineToArgvW()` to avoid loading "shell32.dll" Move "font.h" -> "raster_font.h" EPS/SVG: use new `out_putsf()` func to output floats, avoiding trailing zeroes & locale dependency EPS: simplify "TR" formula SVG: change font from "Helvetica, sans serif" to "OCR-B, monospace"; use single "<path>" instead of multiple "<rect>"s to draw boxes (reduces file size) Add `EMBED_VECTOR_FONT` to `output_options` (CLI `--embedfont`) to enable embedding of font in vector output - currently only for SVG output of EAN/UPC GUI: use "OCR-B" font for EAN/UPC and "Arimo" for all others (was "Helvetica" for both); paint background of screen preview light grey so as contrasts with whitespace and quiet zones EMF: prefix funcs with `emf_`; simplify string `halign` handling large: rename `large_int` -> `large_uint` CODE128/common: move `c128_hrt_cpy_iso8859_1()` to `hrt_cpy_iso8859_1()` and add `ZINT_WARN_HRT_TRUNCATED` warning (for future use) Various symbologies: replace `printf()` with `fputs()` (symbol->debug) QRCODE: better assert(), removing a NOLINT (2 left) CLI: add some more barcode synonyms for DBAR common: various fiddlings CMake: don't include png.c unless ZINT_USE_PNG (avoids clang warning)
2023-06-12 12:25:55 +12:00
fullstream[j + data_cw] = interleaved_ecc[j];
}
2016-02-17 23:52:48 +13:00
if (debug_print) {
printf("\nData Stream (%d): \n", data_cw + ecc_cw);
for (j = 0; j < (data_cw + ecc_cw); j++) {
printf("%2X ", fullstream[j]);
}
EAN/UPC: add quiet zone indicators option (API `output_options` `EANUPC_GUARD_WHITESPACE`, CLI `--guardwhitespace`) (ticket #287) EAN-2/EAN-5: HRT now at top instead of at bottom for standalones, following BWIPP CLI: batch mode: don't close input if stdin EAN/UPC: fix excess 1X to right of add-ons Composites: fix excess whitespace; fix quiet zone calcs to allow for linear shifting CLI: use own (Wine) version of `CommandLineToArgvW()` to avoid loading "shell32.dll" Move "font.h" -> "raster_font.h" EPS/SVG: use new `out_putsf()` func to output floats, avoiding trailing zeroes & locale dependency EPS: simplify "TR" formula SVG: change font from "Helvetica, sans serif" to "OCR-B, monospace"; use single "<path>" instead of multiple "<rect>"s to draw boxes (reduces file size) Add `EMBED_VECTOR_FONT` to `output_options` (CLI `--embedfont`) to enable embedding of font in vector output - currently only for SVG output of EAN/UPC GUI: use "OCR-B" font for EAN/UPC and "Arimo" for all others (was "Helvetica" for both); paint background of screen preview light grey so as contrasts with whitespace and quiet zones EMF: prefix funcs with `emf_`; simplify string `halign` handling large: rename `large_int` -> `large_uint` CODE128/common: move `c128_hrt_cpy_iso8859_1()` to `hrt_cpy_iso8859_1()` and add `ZINT_WARN_HRT_TRUNCATED` warning (for future use) Various symbologies: replace `printf()` with `fputs()` (symbol->debug) QRCODE: better assert(), removing a NOLINT (2 left) CLI: add some more barcode synonyms for DBAR common: various fiddlings CMake: don't include png.c unless ZINT_USE_PNG (avoids clang warning)
2023-06-12 12:25:55 +12:00
fputc('\n', stdout);
}
2016-02-17 23:52:48 +13:00
}
static void qr_place_finder(unsigned char grid[], const int size, const int x, const int y) {
int xp, yp;
2017-08-07 19:37:02 +12:00
char finder[] = {0x7F, 0x41, 0x5D, 0x5D, 0x5D, 0x41, 0x7F};
2017-10-24 08:37:52 +13:00
for (xp = 0; xp < 7; xp++) {
for (yp = 0; yp < 7; yp++) {
2017-08-07 19:37:02 +12:00
if (finder[yp] & 0x40 >> xp) {
grid[((yp + y) * size) + (xp + x)] = 0x11;
} else {
grid[((yp + y) * size) + (xp + x)] = 0x10;
}
}
}
2016-02-17 23:52:48 +13:00
}
static void qr_place_align(unsigned char grid[], const int size, int x, int y) {
int xp, yp;
2017-08-07 19:37:02 +12:00
char alignment[] = {0x1F, 0x11, 0x15, 0x11, 0x1F};
x -= 2;
y -= 2; /* Input values represent centre of pattern */
for (xp = 0; xp < 5; xp++) {
for (yp = 0; yp < 5; yp++) {
2017-08-07 19:37:02 +12:00
if (alignment[yp] & 0x10 >> xp) {
grid[((yp + y) * size) + (xp + x)] = 0x11;
} else {
grid[((yp + y) * size) + (xp + x)] = 0x10;
}
}
}
}
static void qr_setup_grid(unsigned char *grid, const int size, const int version) {
int i, toggle = 1;
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
/* Suppress false positive gcc >= 13 warning (when optimizing only) "writing 1 byte into a region of size 0" */
#if defined(__GNUC__) && !defined(__clang__) && __GNUC__ >= 13
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Wstringop-overflow"
#endif
/* Add timing patterns */
for (i = 0; i < size; i++) {
if (toggle == 1) {
grid[(6 * size) + i] = 0x21;
grid[(i * size) + 6] = 0x21;
toggle = 0;
} else {
grid[(6 * size) + i] = 0x20;
grid[(i * size) + 6] = 0x20;
toggle = 1;
}
}
/* Add finder patterns */
qr_place_finder(grid, size, 0, 0);
qr_place_finder(grid, size, 0, size - 7);
qr_place_finder(grid, size, size - 7, 0);
/* Add separators */
for (i = 0; i < 7; i++) {
grid[(7 * size) + i] = 0x10;
grid[(i * size) + 7] = 0x10;
grid[(7 * size) + (size - 1 - i)] = 0x10;
grid[(i * size) + (size - 8)] = 0x10;
grid[((size - 8) * size) + i] = 0x10;
grid[((size - 1 - i) * size) + 7] = 0x10;
}
grid[(7 * size) + 7] = 0x10;
grid[(7 * size) + (size - 8)] = 0x10;
grid[((size - 8) * size) + 7] = 0x10;
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
#if defined(__GNUC__) && !defined(__clang__) && __GNUC__ >= 13
#pragma GCC diagnostic pop
#endif
/* Add alignment patterns */
if (version != 1) {
/* Version 1 does not have alignment patterns */
2017-09-11 03:03:09 +12:00
int loopsize = qr_align_loopsize[version - 1];
int x, y;
for (x = 0; x < loopsize; x++) {
for (y = 0; y < loopsize; y++) {
2017-09-11 03:03:09 +12:00
int xcoord = qr_table_e1[((version - 2) * 7) + x];
int ycoord = qr_table_e1[((version - 2) * 7) + y];
if (!(grid[(ycoord * size) + xcoord] & 0x10)) {
qr_place_align(grid, size, xcoord, ycoord);
}
}
}
}
/* Reserve space for format information */
for (i = 0; i < 8; i++) {
grid[(8 * size) + i] |= 0x20;
grid[(i * size) + 8] |= 0x20;
grid[(8 * size) + (size - 1 - i)] = 0x20;
grid[((size - 1 - i) * size) + 8] = 0x20;
}
grid[(8 * size) + 8] |= 0x20;
grid[((size - 1 - 7) * size) + 8] = 0x21; /* Dark Module from Figure 25 */
/* Reserve space for version information */
if (version >= 7) {
for (i = 0; i < 6; i++) {
grid[((size - 9) * size) + i] = 0x20;
grid[((size - 10) * size) + i] = 0x20;
grid[((size - 11) * size) + i] = 0x20;
grid[(i * size) + (size - 9)] = 0x20;
grid[(i * size) + (size - 10)] = 0x20;
grid[(i * size) + (size - 11)] = 0x20;
}
}
}
static int qr_cwbit(const unsigned char *fullstream, const int i) {
if (fullstream[(i >> 3)] & (0x80 >> (i & 0x07))) {
return 1;
}
return 0;
}
static void qr_populate_grid(unsigned char *grid, const int h_size, const int v_size, const unsigned char *fullstream,
const int cw) {
const int not_rmqr = v_size == h_size;
const int x_start = h_size - (not_rmqr ? 2 : 3); /* For rMQR allow for righthand vertical timing pattern */
int direction = 1; /* up */
int row = 0; /* right hand side */
2017-09-11 03:03:09 +12:00
int i, n, y;
n = cw * 8;
y = v_size - 1;
i = 0;
while (i < n) {
int x = x_start - (row * 2);
int r = y * h_size;
if ((x < 6) && (not_rmqr))
x--; /* skip over vertical timing pattern */
if (!(grid[r + (x + 1)] & 0xf0)) {
grid[r + (x + 1)] = qr_cwbit(fullstream, i);
i++;
}
if (i < n) {
if (!(grid[r + x] & 0xf0)) {
grid[r + x] = qr_cwbit(fullstream, i);
i++;
}
}
if (direction) {
y--;
if (y == -1) {
/* reached the top */
row++;
y = 0;
direction = 0;
}
} else {
y++;
if (y == v_size) {
/* reached the bottom */
row++;
y = v_size - 1;
direction = 1;
}
}
}
}
#ifdef ZINTLOG
static void append_log(const unsigned char log) {
FILE *file;
if ((file = fopen("zintlog.txt", "a+"))) {
fprintf(file, "%02X", log);
(void) fclose(file);
}
}
static void write_log(const char log[]) {
FILE *file;
if ((file = fopen("zintlog.txt", "a+"))) {
fprintf(file, "%s\n", log); /*writes*/
(void) fclose(file);
}
}
#endif
static int qr_evaluate(unsigned char *local, const int size) {
static const unsigned char h1011101[7] = { 1, 0, 1, 1, 1, 0, 1 };
int x, y, r, k, block;
int result = 0;
char state;
int dark_mods;
double percentage;
int a, b, afterCount, beforeCount;
#ifdef ZINTLOG
int result_b = 0;
char str[15];
#endif
/* Suppresses clang-tidy clang-analyzer-core.UndefinedBinaryOperatorResult warnings */
assert(size > 0);
#ifdef ZINTLOG
/* bitmask output */
for (y = 0; y < size; y++) {
for (x = 0; x < size; x++) {
append_log(local[(y * size) + x]);
}
write_log("");
}
#endif
/* Test 1: Adjacent modules in row/column in same colour */
/* Vertical */
for (x = 0; x < size; x++) {
block = 0;
state = 0;
for (y = 0; y < size; y++) {
if (local[(y * size) + x] == state) {
block++;
} else {
if (block >= 5) {
result += block - 2;
}
block = 1;
state = local[(y * size) + x];
}
}
if (block >= 5) {
result += block - 2;
}
}
/* Horizontal */
dark_mods = 0; /* Count dark mods simultaneously (see Test 4 below) */
for (y = 0; y < size; y++) {
r = y * size;
block = 0;
state = 0;
for (x = 0; x < size; x++) {
if (local[r + x] == state) {
block++;
} else {
if (block >= 5) {
result += block - 2;
}
block = 1;
state = local[r + x];
}
if (state) {
dark_mods++;
}
}
if (block >= 5) {
result += block - 2;
}
}
#ifdef ZINTLOG
/* output Test 1 */
sprintf(str, "%d", result);
result_b = result;
write_log(str);
#endif
/* Test 2: Block of modules in same color */
for (x = 0; x < size - 1; x++) {
for (y = 0; y < size - 1; y++) {
k = local[(y * size) + x];
if (((k == local[((y + 1) * size) + x]) &&
(k == local[(y * size) + (x + 1)])) &&
(k == local[((y + 1) * size) + (x + 1)])) {
result += 3;
}
}
}
#ifdef ZINTLOG
/* output Test 2 */
sprintf(str, "%d", result - result_b);
result_b = result;
write_log(str);
#endif
/* Test 3: 1:1:3:1:1 ratio pattern in row/column */
/* Vertical */
for (x = 0; x < size; x++) {
for (y = 0; y <= (size - 7); y++) {
if (local[y * size + x] && !local[(y + 1) * size + x] && local[(y + 2) * size + x] &&
local[(y + 3) * size + x] && local[(y + 4) * size + x] &&
!local[(y + 5) * size + x] && local[(y + 6) * size + x]) {
/* Pattern found, check before and after */
beforeCount = 0;
for (b = (y - 1); b >= (y - 4); b--) {
if (b < 0) { /* Count < edge as whitespace */
beforeCount = 4;
break;
}
if (local[(b * size) + x]) {
break;
}
beforeCount++;
}
if (beforeCount == 4) {
/* Pattern is preceded by light area 4 modules wide */
result += 40;
} else {
afterCount = 0;
for (a = (y + 7); a <= (y + 10); a++) {
if (a >= size) { /* Count > edge as whitespace */
afterCount = 4;
break;
}
if (local[(a * size) + x]) {
break;
}
afterCount++;
}
if (afterCount == 4) {
/* Pattern is followed by light area 4 modules wide */
result += 40;
}
}
y += 3; /* Skip to next possible match */
}
}
}
/* Horizontal */
for (y = 0; y < size; y++) {
r = y * size;
for (x = 0; x <= (size - 7); x++) {
if (memcmp(local + r + x, h1011101, 7) == 0) {
/* Pattern found, check before and after */
beforeCount = 0;
for (b = (x - 1); b >= (x - 4); b--) {
if (b < 0) { /* Count < edge as whitespace */
beforeCount = 4;
break;
}
if (local[r + b]) {
break;
}
beforeCount++;
}
if (beforeCount == 4) {
/* Pattern is preceded by light area 4 modules wide */
result += 40;
} else {
afterCount = 0;
for (a = (x + 7); a <= (x + 10); a++) {
if (a >= size) { /* Count > edge as whitespace */
afterCount = 4;
break;
}
if (local[r + a]) {
break;
}
afterCount++;
}
if (afterCount == 4) {
/* Pattern is followed by light area 4 modules wide */
result += 40;
}
}
x += 3; /* Skip to next possible match */
}
}
}
#ifdef ZINTLOG
/* output Test 3 */
sprintf(str, "%d", result - result_b);
result_b = result;
write_log(str);
#endif
/* Test 4: Proportion of dark modules in entire symbol */
percentage = (100.0 * dark_mods) / (size * size);
k = (int) (fabs(percentage - 50.0) / 5.0);
result += 10 * k;
#ifdef ZINTLOG
/* output Test 4+summary */
sprintf(str, "%d", result - result_b);
write_log(str);
write_log("==========");
sprintf(str, "%d", result);
write_log(str);
#endif
return result;
}
/* Add format information to grid */
static void qr_add_format_info(unsigned char *grid, const int size, const int ecc_level, const int pattern) {
int format = pattern;
unsigned int seq;
int i;
switch (ecc_level) {
case QR_LEVEL_L: format |= 0x08;
break;
case QR_LEVEL_Q: format |= 0x18;
break;
case QR_LEVEL_H: format |= 0x10;
break;
}
seq = qr_annex_c[format];
for (i = 0; i < 6; i++) {
grid[(i * size) + 8] |= (seq >> i) & 0x01;
}
for (i = 0; i < 8; i++) {
grid[(8 * size) + (size - i - 1)] |= (seq >> i) & 0x01;
}
for (i = 0; i < 6; i++) {
grid[(8 * size) + (5 - i)] |= (seq >> (i + 9)) & 0x01;
}
for (i = 0; i < 7; i++) {
grid[(((size - 7) + i) * size) + 8] |= (seq >> (i + 8)) & 0x01;
}
grid[(7 * size) + 8] |= (seq >> 6) & 0x01;
grid[(8 * size) + 8] |= (seq >> 7) & 0x01;
grid[(8 * size) + 7] |= (seq >> 8) & 0x01;
}
static int qr_apply_bitmask(unsigned char *grid, const int size, const int ecc_level, const int user_mask,
const int fast_encode, const int debug_print) {
int x, y;
int r, k;
int bit;
int pattern, penalty[8];
int best_pattern;
int size_squared = size * size;
unsigned char *mask = (unsigned char *) z_alloca(size_squared);
unsigned char *local = (unsigned char *) z_alloca(size_squared);
#ifdef ZINTLOG
char str[15];
#endif
/* Perform data masking */
memset(mask, 0, size_squared);
for (y = 0; y < size; y++) {
r = y * size;
for (x = 0; x < size; x++) {
/* all eight bitmask variants are encoded in the 8 bits of the bytes that make up the mask array. */
if (!(grid[r + x] & 0xf0)) { /* exclude areas not to be masked. */
if (((y + x) & 1) == 0) {
mask[r + x] |= 0x01;
}
if (!fast_encode) {
if ((y & 1) == 0) {
mask[r + x] |= 0x02;
}
}
if ((x % 3) == 0) {
mask[r + x] |= 0x04;
}
if (!fast_encode) {
if (((y + x) % 3) == 0) {
mask[r + x] |= 0x08;
}
}
if ((((y / 2) + (x / 3)) & 1) == 0) {
mask[r + x] |= 0x10;
}
if (!fast_encode) {
if ((y * x) % 6 == 0) { /* Equivalent to (y * x) % 2 + (y * x) % 3 == 0 */
mask[r + x] |= 0x20;
}
if (((((y * x) & 1) + ((y * x) % 3)) & 1) == 0) {
mask[r + x] |= 0x40;
}
}
if (((((y + x) & 1) + ((y * x) % 3)) & 1) == 0) {
mask[r + x] |= 0x80;
}
}
}
}
if (user_mask) {
best_pattern = user_mask - 1;
} else {
/* all eight bitmask variants have been encoded in the 8 bits of the bytes
* that make up the mask array. select them for evaluation according to the
* desired pattern.*/
best_pattern = 0;
for (pattern = 0; pattern < 8; pattern++) {
if (fast_encode && pattern != 0 && pattern != 2 && pattern != 4 && pattern != 7) {
continue;
}
bit = 1 << pattern;
for (k = 0; k < size_squared; k++) {
if (mask[k] & bit) {
local[k] = grid[k] ^ 0x01;
} else {
local[k] = grid[k] & 0x0f;
}
}
qr_add_format_info(local, size, ecc_level, pattern);
penalty[pattern] = qr_evaluate(local, size);
if (penalty[pattern] < penalty[best_pattern]) {
best_pattern = pattern;
}
}
}
if (debug_print) {
printf("Mask: %d (%s)", best_pattern, user_mask ? "specified" : fast_encode ? "fast automatic": "automatic");
if (!user_mask) {
if (fast_encode) {
printf(" 0:%d 2:%d 4:%d 7:%d", penalty[0], penalty[2], penalty[4], penalty[7]);
} else {
for (pattern = 0; pattern < 8; pattern++) printf(" %d:%d", pattern, penalty[pattern]);
}
}
EAN/UPC: add quiet zone indicators option (API `output_options` `EANUPC_GUARD_WHITESPACE`, CLI `--guardwhitespace`) (ticket #287) EAN-2/EAN-5: HRT now at top instead of at bottom for standalones, following BWIPP CLI: batch mode: don't close input if stdin EAN/UPC: fix excess 1X to right of add-ons Composites: fix excess whitespace; fix quiet zone calcs to allow for linear shifting CLI: use own (Wine) version of `CommandLineToArgvW()` to avoid loading "shell32.dll" Move "font.h" -> "raster_font.h" EPS/SVG: use new `out_putsf()` func to output floats, avoiding trailing zeroes & locale dependency EPS: simplify "TR" formula SVG: change font from "Helvetica, sans serif" to "OCR-B, monospace"; use single "<path>" instead of multiple "<rect>"s to draw boxes (reduces file size) Add `EMBED_VECTOR_FONT` to `output_options` (CLI `--embedfont`) to enable embedding of font in vector output - currently only for SVG output of EAN/UPC GUI: use "OCR-B" font for EAN/UPC and "Arimo" for all others (was "Helvetica" for both); paint background of screen preview light grey so as contrasts with whitespace and quiet zones EMF: prefix funcs with `emf_`; simplify string `halign` handling large: rename `large_int` -> `large_uint` CODE128/common: move `c128_hrt_cpy_iso8859_1()` to `hrt_cpy_iso8859_1()` and add `ZINT_WARN_HRT_TRUNCATED` warning (for future use) Various symbologies: replace `printf()` with `fputs()` (symbol->debug) QRCODE: better assert(), removing a NOLINT (2 left) CLI: add some more barcode synonyms for DBAR common: various fiddlings CMake: don't include png.c unless ZINT_USE_PNG (avoids clang warning)
2023-06-12 12:25:55 +12:00
fputc('\n', stdout);
}
#ifdef ZINTLOG
sprintf(str, "%d", best_pattern);
write_log("chose pattern:");
write_log(str);
#endif
/* Apply mask */
if (!user_mask && best_pattern == 7) { /* Reuse last */
memcpy(grid, local, size_squared);
} else {
bit = 1 << best_pattern;
for (y = 0; y < size_squared; y++) {
if (mask[y] & bit) {
grid[y] ^= 0x01;
}
}
}
return best_pattern;
}
/* Add version information */
static void qr_add_version_info(unsigned char *grid, const int size, const int version) {
int i;
EAN/UPC: add quiet zone indicators option (API `output_options` `EANUPC_GUARD_WHITESPACE`, CLI `--guardwhitespace`) (ticket #287) EAN-2/EAN-5: HRT now at top instead of at bottom for standalones, following BWIPP CLI: batch mode: don't close input if stdin EAN/UPC: fix excess 1X to right of add-ons Composites: fix excess whitespace; fix quiet zone calcs to allow for linear shifting CLI: use own (Wine) version of `CommandLineToArgvW()` to avoid loading "shell32.dll" Move "font.h" -> "raster_font.h" EPS/SVG: use new `out_putsf()` func to output floats, avoiding trailing zeroes & locale dependency EPS: simplify "TR" formula SVG: change font from "Helvetica, sans serif" to "OCR-B, monospace"; use single "<path>" instead of multiple "<rect>"s to draw boxes (reduces file size) Add `EMBED_VECTOR_FONT` to `output_options` (CLI `--embedfont`) to enable embedding of font in vector output - currently only for SVG output of EAN/UPC GUI: use "OCR-B" font for EAN/UPC and "Arimo" for all others (was "Helvetica" for both); paint background of screen preview light grey so as contrasts with whitespace and quiet zones EMF: prefix funcs with `emf_`; simplify string `halign` handling large: rename `large_int` -> `large_uint` CODE128/common: move `c128_hrt_cpy_iso8859_1()` to `hrt_cpy_iso8859_1()` and add `ZINT_WARN_HRT_TRUNCATED` warning (for future use) Various symbologies: replace `printf()` with `fputs()` (symbol->debug) QRCODE: better assert(), removing a NOLINT (2 left) CLI: add some more barcode synonyms for DBAR common: various fiddlings CMake: don't include png.c unless ZINT_USE_PNG (avoids clang warning)
2023-06-12 12:25:55 +12:00
unsigned int version_data = qr_annex_d[version - 7];
for (i = 0; i < 6; i++) {
grid[((size - 11) * size) + i] |= (version_data >> (i * 3)) & 1;
grid[((size - 10) * size) + i] |= (version_data >> ((i * 3) + 1)) & 1;
grid[((size - 9) * size) + i] |= (version_data >> ((i * 3) + 2)) & 1;
grid[(i * size) + (size - 11)] |= (version_data >> (i * 3)) & 1;
grid[(i * size) + (size - 10)] |= (version_data >> ((i * 3) + 1)) & 1;
grid[(i * size) + (size - 9)] |= (version_data >> ((i * 3) + 2)) & 1;
}
}
/* Find the length of the block starting from 'start' */
static int qr_blockLength(const int start, const char mode[], const int length) {
int i;
int count = 0;
char start_mode = mode[start];
i = start;
do {
count++;
} while (((i + count) < length) && (mode[i + count] == start_mode));
return count;
}
/* Calculate the actual bitlength of the proposed binary string */
static int qr_calc_binlen(const int version, char mode[], const unsigned int ddata[], const int length,
const int mode_preset, const int gs1, const int eci, const int debug_print) {
int i, j;
char currentMode;
int count = 0;
int alphalength;
2019-12-17 06:31:52 +13:00
int blocklength;
if (!mode_preset) {
qr_define_mode(mode, ddata, length, gs1, version, debug_print);
}
currentMode = ' '; /* Null */
if (eci != 0) { /* Not applicable to MICROQR */
count += 4;
if (eci <= 127) {
count += 8;
} else if (eci <= 16383) {
count += 16;
} else {
count += 24;
}
2016-08-16 23:43:41 +12:00
}
for (i = 0; i < length; i++) {
if (mode[i] != currentMode) {
count += qr_mode_bits(version) + qr_cci_bits(version, mode[i]);
blocklength = qr_blockLength(i, mode, length);
switch (mode[i]) {
case 'K':
2019-12-17 06:31:52 +13:00
count += (blocklength * 13);
break;
case 'B':
2019-12-17 06:31:52 +13:00
for (j = i; j < (i + blocklength); j++) {
if (ddata[j] > 0xff) {
count += 16;
} else {
count += 8;
}
}
break;
case 'A':
2019-12-17 06:31:52 +13:00
alphalength = blocklength;
if (gs1) {
/* In alphanumeric mode % becomes %% */
2019-12-17 06:31:52 +13:00
for (j = i; j < (i + blocklength); j++) {
if (ddata[j] == '%') {
2019-12-17 06:31:52 +13:00
alphalength++;
}
}
}
switch (alphalength % 2) {
case 0:
count += (alphalength / 2) * 11;
break;
case 1:
count += ((alphalength - 1) / 2) * 11;
count += 6;
break;
}
break;
case 'N':
2019-12-17 06:31:52 +13:00
switch (blocklength % 3) {
case 0:
2019-12-17 06:31:52 +13:00
count += (blocklength / 3) * 10;
break;
case 1:
2019-12-17 06:31:52 +13:00
count += ((blocklength - 1) / 3) * 10;
count += 4;
break;
case 2:
2019-12-17 06:31:52 +13:00
count += ((blocklength - 2) / 3) * 10;
count += 7;
break;
}
break;
}
currentMode = mode[i];
}
}
return count;
}
/* Call `qr_calc_binlen()` on each segment */
static int qr_calc_binlen_segs(const int version, char mode[], const unsigned int ddata[],
const struct zint_seg segs[], const int seg_count, const struct zint_structapp *p_structapp,
const int mode_preset, const int gs1, const int debug_print) {
int i;
int count = 0;
const unsigned int *dd = ddata;
char *m = mode;
if (p_structapp) {
count += 4 + 8 + 8;
}
if (gs1) { /* Not applicable to MICROQR */
if (version < RMQR_VERSION) {
count += 4;
} else {
count += 3;
}
}
for (i = 0; i < seg_count; i++) {
count += qr_calc_binlen(version, m, dd, segs[i].length, mode_preset, gs1, segs[i].eci, debug_print);
m += segs[i].length;
dd += segs[i].length;
}
if (debug_print) {
printf("Estimated Binary Length: %d (version %d, gs1 %d)\n", count, version, gs1);
}
2019-12-17 06:31:52 +13:00
return count;
}
/* Helper to process source data into `ddata` array */
static int qr_prep_data(struct zint_symbol *symbol, struct zint_seg segs[], const int seg_count,
unsigned int ddata[]) {
int warn_number = 0;
int i;
/* If ZINT_FULL_MULTIBYTE use Kanji mode in DATA_MODE or for non-Shift JIS in UNICODE_MODE */
const int full_multibyte = (symbol->option_3 & 0xFF) == ZINT_FULL_MULTIBYTE;
if ((symbol->input_mode & 0x07) == DATA_MODE) {
sjis_cpy_segs(segs, seg_count, ddata, full_multibyte);
} else {
unsigned int *dd = ddata;
for (i = 0; i < seg_count; i++) {
int done = 0;
if (segs[i].eci != 20 || seg_count > 1) { /* Unless ECI 20 (Shift JIS) or have multiple segments */
/* Try other encodings (ECI 0 defaults to ISO/IEC 8859-1) */
int error_number = sjis_utf8_to_eci(segs[i].eci, segs[i].source, &segs[i].length, dd, full_multibyte);
if (error_number == 0) {
done = 1;
} else if (segs[i].eci || seg_count > 1) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxtf(error_number, symbol, 575, "Invalid character in input for ECI '%d'", segs[i].eci);
}
}
if (!done) {
/* Try Shift-JIS */
int error_number = sjis_utf8(symbol, segs[i].source, &segs[i].length, dd);
if (error_number != 0) {
return error_number;
}
if (segs[i].eci != 20) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
warn_number = errtxt(ZINT_WARN_NONCOMPLIANT, symbol, 760,
"Converted to Shift JIS but no ECI specified");
}
}
dd += segs[i].length;
}
}
return warn_number;
}
INTERNAL int qrcode(struct zint_symbol *symbol, struct zint_seg segs[], const int seg_count) {
int warn_number;
int i, j, est_binlen, prev_est_binlen;
int ecc_level, autosize, version, max_cw, target_codewords, blocks, size;
int bitmask;
int user_mask;
int canShrink;
int size_squared;
const struct zint_structapp *p_structapp = NULL;
const int gs1 = ((symbol->input_mode & 0x07) == GS1_MODE);
const int fast_encode = symbol->input_mode & FAST_MODE;
const int debug_print = symbol->debug & ZINT_DEBUG_PRINT;
const int eci_length_segs = get_eci_length_segs(segs, seg_count);
struct zint_seg *local_segs = (struct zint_seg *) z_alloca(sizeof(struct zint_seg) * seg_count);
unsigned int *ddata = (unsigned int *) z_alloca(sizeof(unsigned int) * eci_length_segs);
char *mode = (char *) z_alloca(eci_length_segs);
char *prev_mode = (char *) z_alloca(eci_length_segs);
unsigned char *datastream;
unsigned char *fullstream;
unsigned char *grid;
user_mask = (symbol->option_3 >> 8) & 0x0F; /* User mask is pattern + 1, so >= 1 and <= 8 */
if (user_mask > 8) {
user_mask = 0; /* Ignore */
}
segs_cpy(symbol, segs, seg_count, local_segs); /* Shallow copy (needed to set default ECIs & protect lengths) */
warn_number = qr_prep_data(symbol, local_segs, seg_count, ddata);
if (warn_number >= ZINT_ERROR) {
return warn_number;
}
if (symbol->structapp.count) {
if (symbol->structapp.count < 2 || symbol->structapp.count > 16) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxtf(ZINT_ERROR_INVALID_OPTION, symbol, 750,
"Structured Append count '%d' out of range (2 to 16)", symbol->structapp.count);
}
if (symbol->structapp.index < 1 || symbol->structapp.index > symbol->structapp.count) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxtf(ZINT_ERROR_INVALID_OPTION, symbol, 751,
"Structured Append index '%1$d' out of range (1 to count %2$d)",
symbol->structapp.index, symbol->structapp.count);
}
if (symbol->structapp.id[0]) {
int id, id_len;
- API: add new zint_symbol `dpmm` field for output resolution (BMP/ EMF/PCX/PNG/TIF only, i.e. excluding EPS, GIF & SVG) - Add support for specifying scale by X-dimension and resolution with new option `--scalexdimdp` for CLI/Tcl & new API function `ZBarcode_Scale_From_XdimDp()` (+ `ZBarcode_XdimDp_From_Scale()` & `ZBarcode_Default_Xdim()`) and new GUI popup; manual: document - BMP/EMF/PCX/PNG/TIF: use new `dpmm` resolution field (for EMF following Inkscape) - backend_qt: add `dpmm()`, `vectorWidth()`, `vectorHeight()`, `noPng()`, `getVersion()`, `takesGS1AIData()`, & `XdimDp` stuff incl. new `QZintXdimDp` struct for passing around scale vars & use in `getAsCLI()`; add comments - Raise `scale` limit to 200 (from 100) to allow for large dpmm - output: create directories & subdirectories as necessary for output path using new function `out_fopen()` and use in BMP/EMF/ EPS/GIF/PCX/PNG/SVG/TIF - DPLEIT/DPIDENT: format HRT according to (incomplete) documentation, and set default height to 72X (from 50X) - CODE128B renamed to CODE128AB as can use subsets A and/or B - CODABAR: fix minimum height calc - EMF: fix indexing of handles (zero-based not 1-based) - GUI: fix symbology zap (previous technique of clearing and re-loading settings without doing a sync no longer works); fix UPCEAN guard descent enable - MAILMARK: better error message if input < 14 characters - GUI: add "Default" button for DAFT tracker ratio & enable/disable various default buttons; use new `takesGS1AIData()` to enable/disable GS1-specific checkboxes - CLI: use new `validate_float()` to parse float options (7 significant digits allowed only, no scientific notation) - DATAMATRIX/GRIDMATRIX/PDF417/QR/ULTRA: micro-optimize structapp ID parse - library/CLI: fiddle with static asserts (make CHAR_BIT sensitive, supposedly) - win32/README: update building libpng (assembly removed) - README.linux: document incompatibility of Qt6 >= 6.3 - manual: expand Barcode Studio waffle - test suite: change range separator to hyphen and allow multiple excludes
2022-12-03 10:39:01 +13:00
for (id_len = 1; id_len < 4 && symbol->structapp.id[id_len]; id_len++);
- API: add new zint_symbol `dpmm` field for output resolution (BMP/ EMF/PCX/PNG/TIF only, i.e. excluding EPS, GIF & SVG) - Add support for specifying scale by X-dimension and resolution with new option `--scalexdimdp` for CLI/Tcl & new API function `ZBarcode_Scale_From_XdimDp()` (+ `ZBarcode_XdimDp_From_Scale()` & `ZBarcode_Default_Xdim()`) and new GUI popup; manual: document - BMP/EMF/PCX/PNG/TIF: use new `dpmm` resolution field (for EMF following Inkscape) - backend_qt: add `dpmm()`, `vectorWidth()`, `vectorHeight()`, `noPng()`, `getVersion()`, `takesGS1AIData()`, & `XdimDp` stuff incl. new `QZintXdimDp` struct for passing around scale vars & use in `getAsCLI()`; add comments - Raise `scale` limit to 200 (from 100) to allow for large dpmm - output: create directories & subdirectories as necessary for output path using new function `out_fopen()` and use in BMP/EMF/ EPS/GIF/PCX/PNG/SVG/TIF - DPLEIT/DPIDENT: format HRT according to (incomplete) documentation, and set default height to 72X (from 50X) - CODE128B renamed to CODE128AB as can use subsets A and/or B - CODABAR: fix minimum height calc - EMF: fix indexing of handles (zero-based not 1-based) - GUI: fix symbology zap (previous technique of clearing and re-loading settings without doing a sync no longer works); fix UPCEAN guard descent enable - MAILMARK: better error message if input < 14 characters - GUI: add "Default" button for DAFT tracker ratio & enable/disable various default buttons; use new `takesGS1AIData()` to enable/disable GS1-specific checkboxes - CLI: use new `validate_float()` to parse float options (7 significant digits allowed only, no scientific notation) - DATAMATRIX/GRIDMATRIX/PDF417/QR/ULTRA: micro-optimize structapp ID parse - library/CLI: fiddle with static asserts (make CHAR_BIT sensitive, supposedly) - win32/README: update building libpng (assembly removed) - README.linux: document incompatibility of Qt6 >= 6.3 - manual: expand Barcode Studio waffle - test suite: change range separator to hyphen and allow multiple excludes
2022-12-03 10:39:01 +13:00
if (id_len > 3) { /* Max value 255 */
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxtf(ZINT_ERROR_INVALID_OPTION, symbol, 752,
"Structured Append ID length %d too long (3 digit maximum)", id_len);
}
id = to_int((const unsigned char *) symbol->structapp.id, id_len);
if (id == -1) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxt(ZINT_ERROR_INVALID_OPTION, symbol, 753, "Invalid Structured Append ID (digits only)");
}
if (id > 255) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxtf(ZINT_ERROR_INVALID_OPTION, symbol, 754,
"Structured Append ID value '%d' out of range (0 to 255)", id);
}
}
p_structapp = &symbol->structapp;
}
/* GS1 General Specifications 22.0 section 5.7.3 says Structured Append and ECIs not supported
for GS1 QR Code so check and return ZINT_WARN_NONCOMPLIANT if either true */
if (gs1 && warn_number == 0) {
for (i = 0; i < seg_count; i++) {
if (local_segs[i].eci) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
warn_number = errtxt(ZINT_WARN_NONCOMPLIANT, symbol, 755,
"Using ECI in GS1 mode not supported by GS1 standards");
break;
}
}
if (warn_number == 0 && p_structapp) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
warn_number = errtxt(ZINT_WARN_NONCOMPLIANT, symbol, 756,
"Using Structured Append in GS1 mode not supported by GS1 standards");
}
}
est_binlen = qr_calc_binlen_segs(40, mode, ddata, local_segs, seg_count, p_structapp, 0 /*mode_preset*/, gs1,
debug_print);
if ((symbol->option_1 >= 1) && (symbol->option_1 <= 4)) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
ecc_level = symbol->option_1 - 1;
} else {
ecc_level = QR_LEVEL_L;
}
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
max_cw = qr_data_codewords[ecc_level][39];
if (est_binlen > (8 * max_cw)) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
if (ecc_level == QR_LEVEL_L) {
return errtxtf(ZINT_ERROR_TOO_LONG, symbol, 567, "Input too long, requires %1$d codewords (maximum %2$d)",
(est_binlen + 7) / 8, max_cw);
}
return errtxtf(ZINT_ERROR_TOO_LONG, symbol, 561,
"Input too long for ECC level %1$c, requires %2$d codewords (maximum %3$d)",
qr_ecc_level_names[ecc_level], (est_binlen + 7) / 8, max_cw);
}
autosize = 40;
for (i = 39; i >= 0; i--) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
if ((8 * qr_data_codewords[ecc_level][i]) >= est_binlen) {
autosize = i + 1;
}
}
2019-12-17 06:31:52 +13:00
if (autosize != 40) {
/* Save version 40 estimate in case incorrect costings in `qr_define_mode()` lead to its `mode` being better
than current lower version one */
prev_est_binlen = est_binlen;
est_binlen = qr_calc_binlen_segs(autosize, mode, ddata, local_segs, seg_count, p_structapp, 0 /*mode_preset*/,
gs1, debug_print);
if (prev_est_binlen < est_binlen) { /* Shouldn't happen */
assert(0); /* Not reached (hopefully) */
/* Defensively use version 40 `mode` to avoid crashes (ticket #300) */
est_binlen = qr_calc_binlen_segs(40, mode, ddata, local_segs, seg_count, p_structapp, 0 /*mode_preset*/,
gs1, debug_print);
assert(est_binlen == prev_est_binlen);
}
2019-12-17 06:31:52 +13:00
}
/* Now see if the optimised binary will fit in a smaller symbol. */
canShrink = 1;
do {
if (autosize == 1) {
canShrink = 0;
} else {
prev_est_binlen = est_binlen;
memcpy(prev_mode, mode, eci_length_segs);
est_binlen = qr_calc_binlen_segs(autosize - 1, mode, ddata, local_segs, seg_count, p_structapp,
0 /*mode_preset*/, gs1, debug_print);
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
if ((8 * qr_data_codewords[ecc_level][autosize - 2]) < est_binlen) {
canShrink = 0;
}
if (canShrink == 1) {
/* Optimisation worked - data will fit in a smaller symbol */
autosize--;
} else {
/* Data did not fit in the smaller symbol, revert to original size */
est_binlen = prev_est_binlen;
memcpy(mode, prev_mode, eci_length_segs);
}
}
} while (canShrink == 1);
version = autosize;
if ((symbol->option_2 >= 1) && (symbol->option_2 <= 40)) {
/* If the user has selected a larger symbol than the smallest available,
then use the size the user has selected, and re-optimise for this
symbol size.
*/
if (symbol->option_2 > version) {
version = symbol->option_2;
est_binlen = qr_calc_binlen_segs(symbol->option_2, mode, ddata, local_segs, seg_count, p_structapp,
0 /*mode_preset*/, gs1, debug_print);
}
2017-10-24 08:37:52 +13:00
if (symbol->option_2 < version) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxtf(ZINT_ERROR_TOO_LONG, symbol, 569,
"Input too long for Version %1$d-%2$c, requires %3$d codewords (maximum %4$d)",
symbol->option_2, qr_ecc_level_names[ecc_level], (est_binlen + 7) / 8,
qr_data_codewords[ecc_level][symbol->option_2 - 1]);
}
}
2019-12-17 06:31:52 +13:00
/* Ensure maxium error correction capacity unless user-specified */
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
if (symbol->option_1 == -1 || symbol->option_1 - 1 != ecc_level) {
if (est_binlen <= qr_data_codewords[QR_LEVEL_H][version - 1] * 8) {
ecc_level = QR_LEVEL_H;
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
} else if (est_binlen <= qr_data_codewords[QR_LEVEL_Q][version - 1] * 8) {
ecc_level = QR_LEVEL_Q;
} else if (est_binlen <= qr_data_codewords[QR_LEVEL_M][version - 1] * 8) {
ecc_level = QR_LEVEL_M;
2019-12-17 06:31:52 +13:00
}
}
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
target_codewords = qr_data_codewords[ecc_level][version - 1];
blocks = qr_blocks[ecc_level][version - 1];
if (debug_print) {
printf("Minimum codewords: %d\n", (est_binlen + 7) / 8);
printf("Selected version: %d-%c (%dx%d)\n",
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
version, qr_ecc_level_names[ecc_level], qr_sizes[version - 1], qr_sizes[version - 1]);
printf("Number of data codewords in symbol: %d\n", target_codewords);
printf("Number of ECC blocks: %d\n", blocks);
}
datastream = (unsigned char *) z_alloca(target_codewords + 1);
fullstream = (unsigned char *) z_alloca(qr_total_codewords[version - 1] + 1);
(void) qr_binary_segs(datastream, version, target_codewords, mode, ddata, local_segs, seg_count, p_structapp, gs1,
Performance improvements for linear encoding and raster output - use fixed-length string tables (mostly) instead of (char *) pointer ones (saves ~40K) - re-use C128Table for CODABLOCKF and CODE16K (required removal of Stop character and extra CODE16K-only entry) - use pointer to destination and copy (memcpy/strcpy(), bin_append_posn()) instead of concatenating (strcat()) (mostly) - replace last remaining bin_append()s with bin_append_posn(); bin_append() removed - add length arg to toupper() and expand() (avoids strlen()) - change is_sane() to use table-based flags (avoids an iteration) - rename lookup() to is_sane_lookup() and change to check and return posns and use in pointer to destination loops (avoids strcat()s) - remove special case PHARMA in expand() (dealt with in pharma()) - make #define SILVER/CALCIUM/TECHNETIUM/KRSET etc static strings - replace strchr() -> posn() - CODE128: populate destination once in checksum loop; re-use and export some more routines (c128_set_a/b/c(), c128_put_in_set()) for sharing; prefix defines (SHIFTA -> C128_SHIFTA etc) and existing exported routines - use factor XOR toggle trick in checksum calcs (avoids branch) - raster.c: fill out single 1-pixel row and copy using new draw_bar_line(), copy_bar_line() routines; similarly in buffer_plot compare previous line & copy if same (same technique as used to improve non-half-integer scaling, significant performance increase, (c) codemonkey82); also done for PNG (BMP/GIF/PCX/TIFF not done) - raster/vector/output.c: shorten "output_" prefix -> "out_"; sync vector to other raster changes to try to keep source files similar - 2of5.c: prefix "c25_" JAPANPOST: return error if input data truncated (backward incompatible) DAFT: max chars 50 -> 100 common.c: istwodigit() -> is_twodigit() common.c/emf.c/output.c: use some further stripf()s (MSVC6 float variations) library.c: new check_output_args() helper zint.h: add BARCODE_LAST marker and use in library.c QRCODE: remove a NOLINT (requires clang-tidy-13), one remaining CMake: separate no-optimize from ZINT_DEBUG into new ZINT_NOOPT option
2021-10-21 11:05:30 +13:00
est_binlen, debug_print);
#ifdef ZINT_TEST
2019-12-17 06:31:52 +13:00
if (symbol->debug & ZINT_DEBUG_TEST) debug_test_codeword_dump(symbol, datastream, target_codewords);
#endif
qr_add_ecc(fullstream, datastream, version, target_codewords, blocks, debug_print);
size = qr_sizes[version - 1];
size_squared = size * size;
grid = (unsigned char *) z_alloca(size_squared);
memset(grid, 0, size_squared);
qr_setup_grid(grid, size, version);
qr_populate_grid(grid, size, size, fullstream, qr_total_codewords[version - 1]);
if (version >= 7) {
qr_add_version_info(grid, size, version);
}
bitmask = qr_apply_bitmask(grid, size, ecc_level, user_mask, fast_encode, debug_print);
qr_add_format_info(grid, size, ecc_level, bitmask);
symbol->width = size;
symbol->rows = size;
for (i = 0; i < size; i++) {
int r = i * size;
for (j = 0; j < size; j++) {
if (grid[r + j] & 0x01) {
set_module(symbol, i, j);
}
}
symbol->row_height[i] = 1;
}
symbol->height = size;
return warn_number;
}
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
/* Write terminator, padding & ECC */
static int microqr_end(struct zint_symbol *symbol, char binary_data[], int bp, const int ecc_level,
const int version) {
int i, j;
int bits_left;
unsigned char data_blocks[17];
unsigned char ecc_blocks[15];
rs_t rs;
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
const int terminator_bits = qr_terminator_bits(MICROQR_VERSION + version);
const int bits_total = microqr_data[ecc_level][version][0];
const int data_codewords = microqr_data[ecc_level][version][1];
const int ecc_codewords = microqr_data[ecc_level][version][2];
const int bits_end = version == 0 || version == 2 ? 4 : 8;
/* Add terminator */
Performance improvements for linear encoding and raster output - use fixed-length string tables (mostly) instead of (char *) pointer ones (saves ~40K) - re-use C128Table for CODABLOCKF and CODE16K (required removal of Stop character and extra CODE16K-only entry) - use pointer to destination and copy (memcpy/strcpy(), bin_append_posn()) instead of concatenating (strcat()) (mostly) - replace last remaining bin_append()s with bin_append_posn(); bin_append() removed - add length arg to toupper() and expand() (avoids strlen()) - change is_sane() to use table-based flags (avoids an iteration) - rename lookup() to is_sane_lookup() and change to check and return posns and use in pointer to destination loops (avoids strcat()s) - remove special case PHARMA in expand() (dealt with in pharma()) - make #define SILVER/CALCIUM/TECHNETIUM/KRSET etc static strings - replace strchr() -> posn() - CODE128: populate destination once in checksum loop; re-use and export some more routines (c128_set_a/b/c(), c128_put_in_set()) for sharing; prefix defines (SHIFTA -> C128_SHIFTA etc) and existing exported routines - use factor XOR toggle trick in checksum calcs (avoids branch) - raster.c: fill out single 1-pixel row and copy using new draw_bar_line(), copy_bar_line() routines; similarly in buffer_plot compare previous line & copy if same (same technique as used to improve non-half-integer scaling, significant performance increase, (c) codemonkey82); also done for PNG (BMP/GIF/PCX/TIFF not done) - raster/vector/output.c: shorten "output_" prefix -> "out_"; sync vector to other raster changes to try to keep source files similar - 2of5.c: prefix "c25_" JAPANPOST: return error if input data truncated (backward incompatible) DAFT: max chars 50 -> 100 common.c: istwodigit() -> is_twodigit() common.c/emf.c/output.c: use some further stripf()s (MSVC6 float variations) library.c: new check_output_args() helper zint.h: add BARCODE_LAST marker and use in library.c QRCODE: remove a NOLINT (requires clang-tidy-13), one remaining CMake: separate no-optimize from ZINT_DEBUG into new ZINT_NOOPT option
2021-10-21 11:05:30 +13:00
bits_left = bits_total - bp;
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
if (bits_left <= terminator_bits) {
Performance improvements for linear encoding and raster output - use fixed-length string tables (mostly) instead of (char *) pointer ones (saves ~40K) - re-use C128Table for CODABLOCKF and CODE16K (required removal of Stop character and extra CODE16K-only entry) - use pointer to destination and copy (memcpy/strcpy(), bin_append_posn()) instead of concatenating (strcat()) (mostly) - replace last remaining bin_append()s with bin_append_posn(); bin_append() removed - add length arg to toupper() and expand() (avoids strlen()) - change is_sane() to use table-based flags (avoids an iteration) - rename lookup() to is_sane_lookup() and change to check and return posns and use in pointer to destination loops (avoids strcat()s) - remove special case PHARMA in expand() (dealt with in pharma()) - make #define SILVER/CALCIUM/TECHNETIUM/KRSET etc static strings - replace strchr() -> posn() - CODE128: populate destination once in checksum loop; re-use and export some more routines (c128_set_a/b/c(), c128_put_in_set()) for sharing; prefix defines (SHIFTA -> C128_SHIFTA etc) and existing exported routines - use factor XOR toggle trick in checksum calcs (avoids branch) - raster.c: fill out single 1-pixel row and copy using new draw_bar_line(), copy_bar_line() routines; similarly in buffer_plot compare previous line & copy if same (same technique as used to improve non-half-integer scaling, significant performance increase, (c) codemonkey82); also done for PNG (BMP/GIF/PCX/TIFF not done) - raster/vector/output.c: shorten "output_" prefix -> "out_"; sync vector to other raster changes to try to keep source files similar - 2of5.c: prefix "c25_" JAPANPOST: return error if input data truncated (backward incompatible) DAFT: max chars 50 -> 100 common.c: istwodigit() -> is_twodigit() common.c/emf.c/output.c: use some further stripf()s (MSVC6 float variations) library.c: new check_output_args() helper zint.h: add BARCODE_LAST marker and use in library.c QRCODE: remove a NOLINT (requires clang-tidy-13), one remaining CMake: separate no-optimize from ZINT_DEBUG into new ZINT_NOOPT option
2021-10-21 11:05:30 +13:00
if (bits_left) {
bp = bin_append_posn(0, bits_left, binary_data, bp);
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
bits_left = 0;
}
} else {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
bp = bin_append_posn(0, terminator_bits, binary_data, bp);
bits_left -= terminator_bits;
}
if (symbol->debug & ZINT_DEBUG_PRINT) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
printf("M%d Terminated binary (%d): %.*s (bits_left %d)\n", version + 1, bp, bp, binary_data, bits_left);
}
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
/* Manage last (4-bit) block */
if (bits_end == 4 && bits_left && bits_left <= 4) {
bp = bin_append_posn(0, bits_left, binary_data, bp);
bits_left = 0;
}
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
if (bits_left) {
/* Complete current byte */
Performance improvements for linear encoding and raster output - use fixed-length string tables (mostly) instead of (char *) pointer ones (saves ~40K) - re-use C128Table for CODABLOCKF and CODE16K (required removal of Stop character and extra CODE16K-only entry) - use pointer to destination and copy (memcpy/strcpy(), bin_append_posn()) instead of concatenating (strcat()) (mostly) - replace last remaining bin_append()s with bin_append_posn(); bin_append() removed - add length arg to toupper() and expand() (avoids strlen()) - change is_sane() to use table-based flags (avoids an iteration) - rename lookup() to is_sane_lookup() and change to check and return posns and use in pointer to destination loops (avoids strcat()s) - remove special case PHARMA in expand() (dealt with in pharma()) - make #define SILVER/CALCIUM/TECHNETIUM/KRSET etc static strings - replace strchr() -> posn() - CODE128: populate destination once in checksum loop; re-use and export some more routines (c128_set_a/b/c(), c128_put_in_set()) for sharing; prefix defines (SHIFTA -> C128_SHIFTA etc) and existing exported routines - use factor XOR toggle trick in checksum calcs (avoids branch) - raster.c: fill out single 1-pixel row and copy using new draw_bar_line(), copy_bar_line() routines; similarly in buffer_plot compare previous line & copy if same (same technique as used to improve non-half-integer scaling, significant performance increase, (c) codemonkey82); also done for PNG (BMP/GIF/PCX/TIFF not done) - raster/vector/output.c: shorten "output_" prefix -> "out_"; sync vector to other raster changes to try to keep source files similar - 2of5.c: prefix "c25_" JAPANPOST: return error if input data truncated (backward incompatible) DAFT: max chars 50 -> 100 common.c: istwodigit() -> is_twodigit() common.c/emf.c/output.c: use some further stripf()s (MSVC6 float variations) library.c: new check_output_args() helper zint.h: add BARCODE_LAST marker and use in library.c QRCODE: remove a NOLINT (requires clang-tidy-13), one remaining CMake: separate no-optimize from ZINT_DEBUG into new ZINT_NOOPT option
2021-10-21 11:05:30 +13:00
int remainder = 8 - (bp % 8);
if (remainder != 8) {
bp = bin_append_posn(0, remainder, binary_data, bp);
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
bits_left -= remainder;
}
/* Add padding */
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
if (bits_end == 4 && bits_left > 4) {
bits_left -= 4;
}
remainder = bits_left / 8;
for (i = 0; i < remainder; i++) {
Performance improvements for linear encoding and raster output - use fixed-length string tables (mostly) instead of (char *) pointer ones (saves ~40K) - re-use C128Table for CODABLOCKF and CODE16K (required removal of Stop character and extra CODE16K-only entry) - use pointer to destination and copy (memcpy/strcpy(), bin_append_posn()) instead of concatenating (strcat()) (mostly) - replace last remaining bin_append()s with bin_append_posn(); bin_append() removed - add length arg to toupper() and expand() (avoids strlen()) - change is_sane() to use table-based flags (avoids an iteration) - rename lookup() to is_sane_lookup() and change to check and return posns and use in pointer to destination loops (avoids strcat()s) - remove special case PHARMA in expand() (dealt with in pharma()) - make #define SILVER/CALCIUM/TECHNETIUM/KRSET etc static strings - replace strchr() -> posn() - CODE128: populate destination once in checksum loop; re-use and export some more routines (c128_set_a/b/c(), c128_put_in_set()) for sharing; prefix defines (SHIFTA -> C128_SHIFTA etc) and existing exported routines - use factor XOR toggle trick in checksum calcs (avoids branch) - raster.c: fill out single 1-pixel row and copy using new draw_bar_line(), copy_bar_line() routines; similarly in buffer_plot compare previous line & copy if same (same technique as used to improve non-half-integer scaling, significant performance increase, (c) codemonkey82); also done for PNG (BMP/GIF/PCX/TIFF not done) - raster/vector/output.c: shorten "output_" prefix -> "out_"; sync vector to other raster changes to try to keep source files similar - 2of5.c: prefix "c25_" JAPANPOST: return error if input data truncated (backward incompatible) DAFT: max chars 50 -> 100 common.c: istwodigit() -> is_twodigit() common.c/emf.c/output.c: use some further stripf()s (MSVC6 float variations) library.c: new check_output_args() helper zint.h: add BARCODE_LAST marker and use in library.c QRCODE: remove a NOLINT (requires clang-tidy-13), one remaining CMake: separate no-optimize from ZINT_DEBUG into new ZINT_NOOPT option
2021-10-21 11:05:30 +13:00
bp = bin_append_posn(i & 1 ? 0x11 : 0xEC, 8, binary_data, bp);
}
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
if (bits_end == 4) {
bp = bin_append_posn(0, 4, binary_data, bp);
}
}
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
assert((bp & 0x07) == 8 - bits_end);
/* Copy data into codewords */
for (i = 0; i < data_codewords; i++) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
const int bits = i + 1 == data_codewords ? bits_end : 8;
data_blocks[i] = 0;
2017-10-24 08:37:52 +13:00
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
for (j = 0; j < bits; j++) {
if (binary_data[(i * 8) + j] == '1') {
data_blocks[i] |= 0x80 >> j;
}
}
}
#ifdef ZINT_TEST
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
if (symbol->debug & ZINT_DEBUG_TEST) {
char bp_buf[10];
debug_test_codeword_dump(symbol, data_blocks, data_codewords);
sprintf(bp_buf, "%d", bp); /* Append `bp` to detect padding errors */
errtxt_adj(0, symbol, "%s (%s)", bp_buf);
}
#endif
/* Calculate Reed-Solomon error codewords */
rs_init_gf(&rs, 0x11d);
rs_init_code(&rs, ecc_codewords, 0);
rs_encode(&rs, data_codewords, data_blocks, ecc_blocks);
/* Add Reed-Solomon codewords to binary data */
for (i = 0; i < ecc_codewords; i++) {
Performance improvements for linear encoding and raster output - use fixed-length string tables (mostly) instead of (char *) pointer ones (saves ~40K) - re-use C128Table for CODABLOCKF and CODE16K (required removal of Stop character and extra CODE16K-only entry) - use pointer to destination and copy (memcpy/strcpy(), bin_append_posn()) instead of concatenating (strcat()) (mostly) - replace last remaining bin_append()s with bin_append_posn(); bin_append() removed - add length arg to toupper() and expand() (avoids strlen()) - change is_sane() to use table-based flags (avoids an iteration) - rename lookup() to is_sane_lookup() and change to check and return posns and use in pointer to destination loops (avoids strcat()s) - remove special case PHARMA in expand() (dealt with in pharma()) - make #define SILVER/CALCIUM/TECHNETIUM/KRSET etc static strings - replace strchr() -> posn() - CODE128: populate destination once in checksum loop; re-use and export some more routines (c128_set_a/b/c(), c128_put_in_set()) for sharing; prefix defines (SHIFTA -> C128_SHIFTA etc) and existing exported routines - use factor XOR toggle trick in checksum calcs (avoids branch) - raster.c: fill out single 1-pixel row and copy using new draw_bar_line(), copy_bar_line() routines; similarly in buffer_plot compare previous line & copy if same (same technique as used to improve non-half-integer scaling, significant performance increase, (c) codemonkey82); also done for PNG (BMP/GIF/PCX/TIFF not done) - raster/vector/output.c: shorten "output_" prefix -> "out_"; sync vector to other raster changes to try to keep source files similar - 2of5.c: prefix "c25_" JAPANPOST: return error if input data truncated (backward incompatible) DAFT: max chars 50 -> 100 common.c: istwodigit() -> is_twodigit() common.c/emf.c/output.c: use some further stripf()s (MSVC6 float variations) library.c: new check_output_args() helper zint.h: add BARCODE_LAST marker and use in library.c QRCODE: remove a NOLINT (requires clang-tidy-13), one remaining CMake: separate no-optimize from ZINT_DEBUG into new ZINT_NOOPT option
2021-10-21 11:05:30 +13:00
bp = bin_append_posn(ecc_blocks[ecc_codewords - i - 1], 8, binary_data, bp);
}
Performance improvements for linear encoding and raster output - use fixed-length string tables (mostly) instead of (char *) pointer ones (saves ~40K) - re-use C128Table for CODABLOCKF and CODE16K (required removal of Stop character and extra CODE16K-only entry) - use pointer to destination and copy (memcpy/strcpy(), bin_append_posn()) instead of concatenating (strcat()) (mostly) - replace last remaining bin_append()s with bin_append_posn(); bin_append() removed - add length arg to toupper() and expand() (avoids strlen()) - change is_sane() to use table-based flags (avoids an iteration) - rename lookup() to is_sane_lookup() and change to check and return posns and use in pointer to destination loops (avoids strcat()s) - remove special case PHARMA in expand() (dealt with in pharma()) - make #define SILVER/CALCIUM/TECHNETIUM/KRSET etc static strings - replace strchr() -> posn() - CODE128: populate destination once in checksum loop; re-use and export some more routines (c128_set_a/b/c(), c128_put_in_set()) for sharing; prefix defines (SHIFTA -> C128_SHIFTA etc) and existing exported routines - use factor XOR toggle trick in checksum calcs (avoids branch) - raster.c: fill out single 1-pixel row and copy using new draw_bar_line(), copy_bar_line() routines; similarly in buffer_plot compare previous line & copy if same (same technique as used to improve non-half-integer scaling, significant performance increase, (c) codemonkey82); also done for PNG (BMP/GIF/PCX/TIFF not done) - raster/vector/output.c: shorten "output_" prefix -> "out_"; sync vector to other raster changes to try to keep source files similar - 2of5.c: prefix "c25_" JAPANPOST: return error if input data truncated (backward incompatible) DAFT: max chars 50 -> 100 common.c: istwodigit() -> is_twodigit() common.c/emf.c/output.c: use some further stripf()s (MSVC6 float variations) library.c: new check_output_args() helper zint.h: add BARCODE_LAST marker and use in library.c QRCODE: remove a NOLINT (requires clang-tidy-13), one remaining CMake: separate no-optimize from ZINT_DEBUG into new ZINT_NOOPT option
2021-10-21 11:05:30 +13:00
return bp;
}
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
static void microqr_setup_grid(unsigned char *grid, const int size) {
int i, toggle = 1;
/* Add timing patterns */
for (i = 0; i < size; i++) {
if (toggle == 1) {
grid[i] = 0x21;
grid[(i * size)] = 0x21;
toggle = 0;
} else {
grid[i] = 0x20;
grid[(i * size)] = 0x20;
toggle = 1;
}
}
/* Add finder patterns */
qr_place_finder(grid, size, 0, 0);
/* Add separators */
for (i = 0; i < 7; i++) {
grid[(7 * size) + i] = 0x10;
grid[(i * size) + 7] = 0x10;
}
grid[(7 * size) + 7] = 0x10;
/* Reserve space for format information */
for (i = 0; i < 8; i++) {
grid[(8 * size) + i] |= 0x20;
grid[(i * size) + 8] |= 0x20;
}
grid[(8 * size) + 8] |= 20;
}
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
static void microqr_populate_grid(unsigned char *grid, const int size, const char full_stream[], int bp) {
int direction = 1; /* up */
int row = 0; /* right hand side */
Performance improvements for linear encoding and raster output - use fixed-length string tables (mostly) instead of (char *) pointer ones (saves ~40K) - re-use C128Table for CODABLOCKF and CODE16K (required removal of Stop character and extra CODE16K-only entry) - use pointer to destination and copy (memcpy/strcpy(), bin_append_posn()) instead of concatenating (strcat()) (mostly) - replace last remaining bin_append()s with bin_append_posn(); bin_append() removed - add length arg to toupper() and expand() (avoids strlen()) - change is_sane() to use table-based flags (avoids an iteration) - rename lookup() to is_sane_lookup() and change to check and return posns and use in pointer to destination loops (avoids strcat()s) - remove special case PHARMA in expand() (dealt with in pharma()) - make #define SILVER/CALCIUM/TECHNETIUM/KRSET etc static strings - replace strchr() -> posn() - CODE128: populate destination once in checksum loop; re-use and export some more routines (c128_set_a/b/c(), c128_put_in_set()) for sharing; prefix defines (SHIFTA -> C128_SHIFTA etc) and existing exported routines - use factor XOR toggle trick in checksum calcs (avoids branch) - raster.c: fill out single 1-pixel row and copy using new draw_bar_line(), copy_bar_line() routines; similarly in buffer_plot compare previous line & copy if same (same technique as used to improve non-half-integer scaling, significant performance increase, (c) codemonkey82); also done for PNG (BMP/GIF/PCX/TIFF not done) - raster/vector/output.c: shorten "output_" prefix -> "out_"; sync vector to other raster changes to try to keep source files similar - 2of5.c: prefix "c25_" JAPANPOST: return error if input data truncated (backward incompatible) DAFT: max chars 50 -> 100 common.c: istwodigit() -> is_twodigit() common.c/emf.c/output.c: use some further stripf()s (MSVC6 float variations) library.c: new check_output_args() helper zint.h: add BARCODE_LAST marker and use in library.c QRCODE: remove a NOLINT (requires clang-tidy-13), one remaining CMake: separate no-optimize from ZINT_DEBUG into new ZINT_NOOPT option
2021-10-21 11:05:30 +13:00
int i;
int y;
y = size - 1;
i = 0;
do {
2017-09-11 03:03:09 +12:00
int x = (size - 2) - (row * 2);
if (!(grid[(y * size) + (x + 1)] & 0xf0)) {
if (full_stream[i] == '1') {
grid[(y * size) + (x + 1)] = 0x01;
} else {
grid[(y * size) + (x + 1)] = 0x00;
}
i++;
}
Performance improvements for linear encoding and raster output - use fixed-length string tables (mostly) instead of (char *) pointer ones (saves ~40K) - re-use C128Table for CODABLOCKF and CODE16K (required removal of Stop character and extra CODE16K-only entry) - use pointer to destination and copy (memcpy/strcpy(), bin_append_posn()) instead of concatenating (strcat()) (mostly) - replace last remaining bin_append()s with bin_append_posn(); bin_append() removed - add length arg to toupper() and expand() (avoids strlen()) - change is_sane() to use table-based flags (avoids an iteration) - rename lookup() to is_sane_lookup() and change to check and return posns and use in pointer to destination loops (avoids strcat()s) - remove special case PHARMA in expand() (dealt with in pharma()) - make #define SILVER/CALCIUM/TECHNETIUM/KRSET etc static strings - replace strchr() -> posn() - CODE128: populate destination once in checksum loop; re-use and export some more routines (c128_set_a/b/c(), c128_put_in_set()) for sharing; prefix defines (SHIFTA -> C128_SHIFTA etc) and existing exported routines - use factor XOR toggle trick in checksum calcs (avoids branch) - raster.c: fill out single 1-pixel row and copy using new draw_bar_line(), copy_bar_line() routines; similarly in buffer_plot compare previous line & copy if same (same technique as used to improve non-half-integer scaling, significant performance increase, (c) codemonkey82); also done for PNG (BMP/GIF/PCX/TIFF not done) - raster/vector/output.c: shorten "output_" prefix -> "out_"; sync vector to other raster changes to try to keep source files similar - 2of5.c: prefix "c25_" JAPANPOST: return error if input data truncated (backward incompatible) DAFT: max chars 50 -> 100 common.c: istwodigit() -> is_twodigit() common.c/emf.c/output.c: use some further stripf()s (MSVC6 float variations) library.c: new check_output_args() helper zint.h: add BARCODE_LAST marker and use in library.c QRCODE: remove a NOLINT (requires clang-tidy-13), one remaining CMake: separate no-optimize from ZINT_DEBUG into new ZINT_NOOPT option
2021-10-21 11:05:30 +13:00
if (i < bp) {
if (!(grid[(y * size) + x] & 0xf0)) {
if (full_stream[i] == '1') {
grid[(y * size) + x] = 0x01;
} else {
grid[(y * size) + x] = 0x00;
}
i++;
}
}
if (direction) {
y--;
} else {
y++;
}
if (y == 0) {
/* reached the top */
row++;
y = 1;
direction = 0;
}
if (y == size) {
/* reached the bottom */
row++;
y = size - 1;
direction = 1;
}
Performance improvements for linear encoding and raster output - use fixed-length string tables (mostly) instead of (char *) pointer ones (saves ~40K) - re-use C128Table for CODABLOCKF and CODE16K (required removal of Stop character and extra CODE16K-only entry) - use pointer to destination and copy (memcpy/strcpy(), bin_append_posn()) instead of concatenating (strcat()) (mostly) - replace last remaining bin_append()s with bin_append_posn(); bin_append() removed - add length arg to toupper() and expand() (avoids strlen()) - change is_sane() to use table-based flags (avoids an iteration) - rename lookup() to is_sane_lookup() and change to check and return posns and use in pointer to destination loops (avoids strcat()s) - remove special case PHARMA in expand() (dealt with in pharma()) - make #define SILVER/CALCIUM/TECHNETIUM/KRSET etc static strings - replace strchr() -> posn() - CODE128: populate destination once in checksum loop; re-use and export some more routines (c128_set_a/b/c(), c128_put_in_set()) for sharing; prefix defines (SHIFTA -> C128_SHIFTA etc) and existing exported routines - use factor XOR toggle trick in checksum calcs (avoids branch) - raster.c: fill out single 1-pixel row and copy using new draw_bar_line(), copy_bar_line() routines; similarly in buffer_plot compare previous line & copy if same (same technique as used to improve non-half-integer scaling, significant performance increase, (c) codemonkey82); also done for PNG (BMP/GIF/PCX/TIFF not done) - raster/vector/output.c: shorten "output_" prefix -> "out_"; sync vector to other raster changes to try to keep source files similar - 2of5.c: prefix "c25_" JAPANPOST: return error if input data truncated (backward incompatible) DAFT: max chars 50 -> 100 common.c: istwodigit() -> is_twodigit() common.c/emf.c/output.c: use some further stripf()s (MSVC6 float variations) library.c: new check_output_args() helper zint.h: add BARCODE_LAST marker and use in library.c QRCODE: remove a NOLINT (requires clang-tidy-13), one remaining CMake: separate no-optimize from ZINT_DEBUG into new ZINT_NOOPT option
2021-10-21 11:05:30 +13:00
} while (i < bp);
}
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
static int microqr_evaluate(const unsigned char *grid, const int size, const int pattern) {
int sum1, sum2, i, filter = 0, retval;
switch (pattern) {
case 0: filter = 0x01;
break;
case 1: filter = 0x02;
break;
case 2: filter = 0x04;
break;
case 3: filter = 0x08;
break;
}
sum1 = 0;
sum2 = 0;
for (i = 1; i < size; i++) {
if (grid[(i * size) + size - 1] & filter) {
sum1++;
}
if (grid[((size - 1) * size) + i] & filter) {
sum2++;
}
}
if (sum1 <= sum2) {
retval = (sum1 * 16) + sum2;
} else {
retval = (sum2 * 16) + sum1;
}
return retval;
}
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
static int microqr_apply_bitmask(unsigned char *grid, const int size, const int user_mask, const int debug_print) {
int x, y;
int r, k;
int bit;
int pattern, value[4];
int best_pattern;
int size_squared = size * size;
unsigned char *mask = (unsigned char *) z_alloca(size_squared);
unsigned char *eval = (unsigned char *) z_alloca(size_squared);
/* Perform data masking */
memset(mask, 0, size_squared);
for (y = 0; y < size; y++) {
r = y * size;
for (x = 0; x < size; x++) {
if (!(grid[r + x] & 0xf0)) {
if ((y & 1) == 0) {
mask[r + x] |= 0x01;
}
if ((((y / 2) + (x / 3)) & 1) == 0) {
mask[r + x] |= 0x02;
}
if (((((y * x) & 1) + ((y * x) % 3)) & 1) == 0) {
mask[r + x] |= 0x04;
}
if (((((y + x) & 1) + ((y * x) % 3)) & 1) == 0) {
mask[r + x] |= 0x08;
}
}
}
}
if (user_mask) {
best_pattern = user_mask - 1;
} else {
for (k = 0; k < size_squared; k++) {
if (grid[k] & 0x01) {
eval[k] = mask[k] ^ 0xff;
} else {
eval[k] = mask[k];
}
}
/* Evaluate result */
best_pattern = 0;
for (pattern = 0; pattern < 4; pattern++) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
value[pattern] = microqr_evaluate(eval, size, pattern);
if (value[pattern] > value[best_pattern]) {
best_pattern = pattern;
}
}
}
if (debug_print) {
printf("Mask: %d (%s)", best_pattern, user_mask ? "specified" : "automatic");
if (!user_mask) {
for (pattern = 0; pattern < 4; pattern++) printf(" %d:%d", pattern, value[pattern]);
}
EAN/UPC: add quiet zone indicators option (API `output_options` `EANUPC_GUARD_WHITESPACE`, CLI `--guardwhitespace`) (ticket #287) EAN-2/EAN-5: HRT now at top instead of at bottom for standalones, following BWIPP CLI: batch mode: don't close input if stdin EAN/UPC: fix excess 1X to right of add-ons Composites: fix excess whitespace; fix quiet zone calcs to allow for linear shifting CLI: use own (Wine) version of `CommandLineToArgvW()` to avoid loading "shell32.dll" Move "font.h" -> "raster_font.h" EPS/SVG: use new `out_putsf()` func to output floats, avoiding trailing zeroes & locale dependency EPS: simplify "TR" formula SVG: change font from "Helvetica, sans serif" to "OCR-B, monospace"; use single "<path>" instead of multiple "<rect>"s to draw boxes (reduces file size) Add `EMBED_VECTOR_FONT` to `output_options` (CLI `--embedfont`) to enable embedding of font in vector output - currently only for SVG output of EAN/UPC GUI: use "OCR-B" font for EAN/UPC and "Arimo" for all others (was "Helvetica" for both); paint background of screen preview light grey so as contrasts with whitespace and quiet zones EMF: prefix funcs with `emf_`; simplify string `halign` handling large: rename `large_int` -> `large_uint` CODE128/common: move `c128_hrt_cpy_iso8859_1()` to `hrt_cpy_iso8859_1()` and add `ZINT_WARN_HRT_TRUNCATED` warning (for future use) Various symbologies: replace `printf()` with `fputs()` (symbol->debug) QRCODE: better assert(), removing a NOLINT (2 left) CLI: add some more barcode synonyms for DBAR common: various fiddlings CMake: don't include png.c unless ZINT_USE_PNG (avoids clang warning)
2023-06-12 12:25:55 +12:00
fputc('\n', stdout);
}
/* Apply mask */
bit = 1 << best_pattern;
for (k = 0; k < size_squared; k++) {
if (mask[k] & bit) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
grid[k] ^= 0x01;
}
}
return best_pattern;
}
INTERNAL int microqr(struct zint_symbol *symbol, unsigned char source[], int length) {
int i, size, j;
char full_stream[200];
Performance improvements for linear encoding and raster output - use fixed-length string tables (mostly) instead of (char *) pointer ones (saves ~40K) - re-use C128Table for CODABLOCKF and CODE16K (required removal of Stop character and extra CODE16K-only entry) - use pointer to destination and copy (memcpy/strcpy(), bin_append_posn()) instead of concatenating (strcat()) (mostly) - replace last remaining bin_append()s with bin_append_posn(); bin_append() removed - add length arg to toupper() and expand() (avoids strlen()) - change is_sane() to use table-based flags (avoids an iteration) - rename lookup() to is_sane_lookup() and change to check and return posns and use in pointer to destination loops (avoids strcat()s) - remove special case PHARMA in expand() (dealt with in pharma()) - make #define SILVER/CALCIUM/TECHNETIUM/KRSET etc static strings - replace strchr() -> posn() - CODE128: populate destination once in checksum loop; re-use and export some more routines (c128_set_a/b/c(), c128_put_in_set()) for sharing; prefix defines (SHIFTA -> C128_SHIFTA etc) and existing exported routines - use factor XOR toggle trick in checksum calcs (avoids branch) - raster.c: fill out single 1-pixel row and copy using new draw_bar_line(), copy_bar_line() routines; similarly in buffer_plot compare previous line & copy if same (same technique as used to improve non-half-integer scaling, significant performance increase, (c) codemonkey82); also done for PNG (BMP/GIF/PCX/TIFF not done) - raster/vector/output.c: shorten "output_" prefix -> "out_"; sync vector to other raster changes to try to keep source files similar - 2of5.c: prefix "c25_" JAPANPOST: return error if input data truncated (backward incompatible) DAFT: max chars 50 -> 100 common.c: istwodigit() -> is_twodigit() common.c/emf.c/output.c: use some further stripf()s (MSVC6 float variations) library.c: new check_output_args() helper zint.h: add BARCODE_LAST marker and use in library.c QRCODE: remove a NOLINT (requires clang-tidy-13), one remaining CMake: separate no-optimize from ZINT_DEBUG into new ZINT_NOOPT option
2021-10-21 11:05:30 +13:00
int bp;
int full_multibyte;
int user_mask;
unsigned int ddata[40];
char mode[40];
2019-12-17 06:31:52 +13:00
int alpha_used = 0, byte_or_kanji_used = 0;
int version_valid[4];
int binary_count[4];
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
int ecc_level, version;
2019-12-17 06:31:52 +13:00
int bitmask, format, format_full;
int size_squared;
struct zint_seg segs[1];
const int seg_count = 1;
const int debug_print = symbol->debug & ZINT_DEBUG_PRINT;
unsigned char *grid;
if (length > 35) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxtf(ZINT_ERROR_TOO_LONG, symbol, 562, "Input length %d too long (maximum 35)", length);
}
2016-02-17 23:52:48 +13:00
2019-09-02 22:43:14 +12:00
/* Check option 1 in combination with option 2 */
ecc_level = QR_LEVEL_L;
2019-09-02 22:43:14 +12:00
if (symbol->option_1 >= 1 && symbol->option_1 <= 4) {
if (symbol->option_1 == 4) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxt(ZINT_ERROR_INVALID_OPTION, symbol, 566, "Error correction level H not available");
2019-09-02 22:43:14 +12:00
}
if (symbol->option_2 >= 1 && symbol->option_2 <= 4) {
if (symbol->option_2 == 1 && symbol->option_1 != 1) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxt(ZINT_ERROR_INVALID_OPTION, symbol, 574,
"Version M1 supports error correction level L only");
2019-09-02 22:43:14 +12:00
}
if (symbol->option_2 != 4 && symbol->option_1 == 3) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxt(ZINT_ERROR_INVALID_OPTION, symbol, 563, "Error correction level Q requires Version M4");
2019-09-02 22:43:14 +12:00
}
}
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
ecc_level = symbol->option_1 - 1;
2019-09-02 22:43:14 +12:00
}
/* If ZINT_FULL_MULTIBYTE use Kanji mode in DATA_MODE or for non-Shift JIS in UNICODE_MODE */
full_multibyte = (symbol->option_3 & 0xFF) == ZINT_FULL_MULTIBYTE;
user_mask = (symbol->option_3 >> 8) & 0x0F; /* User mask is pattern + 1, so >= 1 and <= 4 */
if (user_mask > 4) {
user_mask = 0; /* Ignore */
}
if ((symbol->input_mode & 0x07) == DATA_MODE) {
sjis_cpy(source, &length, ddata, full_multibyte);
2016-08-16 23:43:41 +12:00
} else {
/* Try ISO 8859-1 conversion first */
int error_number = sjis_utf8_to_eci(3, source, &length, ddata, full_multibyte);
2016-08-16 23:43:41 +12:00
if (error_number != 0) {
/* Try Shift-JIS */
error_number = sjis_utf8(symbol, source, &length, ddata);
if (error_number != 0) {
return error_number;
}
2016-08-16 23:43:41 +12:00
}
}
2019-12-17 06:31:52 +13:00
/* Determine if alpha (excluding numerics), byte or kanji used */
for (i = 0; i < length && (alpha_used == 0 || byte_or_kanji_used == 0); i++) {
if (!z_isdigit(ddata[i])) {
if (qr_is_alpha(ddata[i], 0 /*gs1*/)) {
2019-12-17 06:31:52 +13:00
alpha_used = 1;
} else {
byte_or_kanji_used = 1;
}
}
}
2016-02-17 23:52:48 +13:00
2019-12-17 06:31:52 +13:00
for (i = 0; i < 4; i++) {
version_valid[i] = 1;
}
2016-02-17 23:52:48 +13:00
/* Eliminate possible versions depending on type of content */
2019-12-17 06:31:52 +13:00
if (byte_or_kanji_used) {
version_valid[0] = 0;
version_valid[1] = 0;
2019-12-17 06:31:52 +13:00
} else if (alpha_used) {
version_valid[0] = 0;
}
2016-02-17 23:52:48 +13:00
2019-12-17 06:31:52 +13:00
/* Eliminate possible versions depending on error correction level specified */
if (ecc_level == QR_LEVEL_Q) {
version_valid[0] = 0;
version_valid[1] = 0;
2019-12-17 06:31:52 +13:00
version_valid[2] = 0;
} else if (ecc_level == QR_LEVEL_M) {
2019-12-17 06:31:52 +13:00
version_valid[0] = 0;
}
segs[0].source = source;
segs[0].length = length;
segs[0].eci = 0;
2019-12-17 06:31:52 +13:00
/* Determine length of binary data */
for (i = 0; i < 4; i++) {
if (version_valid[i]) {
binary_count[i] = qr_calc_binlen_segs(MICROQR_VERSION + i, mode, ddata, segs, seg_count,
NULL /*p_structapp*/, 0 /*mode_preset*/, 0 /*gs1*/, debug_print);
2019-12-17 06:31:52 +13:00
} else {
binary_count[i] = 128 + 1;
}
}
2016-02-17 23:52:48 +13:00
2019-12-17 06:31:52 +13:00
/* Eliminate possible versions depending on binary length and error correction level specified */
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
if (binary_count[3] > microqr_data[ecc_level][3][0]) {
return errtxtf(ZINT_ERROR_TOO_LONG, symbol, 565,
"Input too long for Version M4-%1$c, requires %2$d codewords (maximum %3$d)",
qr_ecc_level_names[ecc_level], (binary_count[3] + 7) / 8, microqr_data[ecc_level][3][1]);
}
for (i = 0; i < 3; i++) {
if (binary_count[i] > microqr_data[ecc_level][i][0]) {
version_valid[i] = 0;
}
}
2016-02-17 23:52:48 +13:00
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
/* Auto-select lowest valid size */
version = 3;
if (version_valid[2]) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
version = 2;
}
if (version_valid[1]) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
version = 1;
}
if (version_valid[0]) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
version = 0;
}
2016-02-17 23:52:48 +13:00
/* Get version from user */
if ((symbol->option_2 >= 1) && (symbol->option_2 <= 4)) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
if (symbol->option_2 == 1 && (i = not_sane(NEON_F, source, length))) {
return errtxtf(ZINT_ERROR_INVALID_DATA, symbol, 758,
"Invalid character at position %d in input for Version M1 (digits only)", i);
} else if (symbol->option_2 == 2 && not_sane(QR_ALPHA, source, length)) {
return errtxt(ZINT_ERROR_INVALID_DATA, symbol, 759,
"Invalid character in input for Version M2 (digits, A-Z, space and \"$%*+-./:\" only)");
}
if (symbol->option_2 - 1 >= version) {
2019-09-02 22:43:14 +12:00
version = symbol->option_2 - 1;
} else {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxtf(ZINT_ERROR_TOO_LONG, symbol, 570,
"Input too long for Version M%1$d-%2$c, requires %3$d codewords (maximum %4$d)",
symbol->option_2, qr_ecc_level_names[ecc_level], (binary_count[version] + 7) / 8,
microqr_data[ecc_level][symbol->option_2 - 1][1]);
}
}
2016-02-17 23:52:48 +13:00
2019-12-17 06:31:52 +13:00
/* If there is enough unused space then increase the error correction level, unless user-specified */
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
if (version && (symbol->option_1 == -1 || symbol->option_1 - 1 != ecc_level)) {
if (binary_count[version] <= microqr_data[QR_LEVEL_Q][version][0]) {
ecc_level = QR_LEVEL_Q;
} else if (binary_count[version] <= microqr_data[QR_LEVEL_M][version][0]) {
ecc_level = QR_LEVEL_M;
}
}
2016-02-17 23:52:48 +13:00
qr_define_mode(mode, ddata, length, 0 /*gs1*/, MICROQR_VERSION + version, debug_print);
2016-02-17 23:52:48 +13:00
bp = qr_binary_segs((unsigned char *) full_stream, MICROQR_VERSION + version, 0 /*target_codewords*/, mode, ddata,
segs, seg_count, NULL /*p_structapp*/, 0 /*gs1*/, binary_count[version], debug_print);
if (debug_print) printf("Binary (%d): %.*s\n", bp, bp, full_stream);
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
bp = microqr_end(symbol, full_stream, bp, ecc_level, version);
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
size = microqr_sizes[version];
size_squared = size * size;
grid = (unsigned char *) z_alloca(size_squared);
memset(grid, 0, size_squared);
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
microqr_setup_grid(grid, size);
microqr_populate_grid(grid, size, full_stream, bp);
bitmask = microqr_apply_bitmask(grid, size, user_mask, debug_print);
/* Add format data */
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
format = version ? (version - 1) * 2 + ecc_level + 1 : 0;
if (debug_print) {
printf("Version: M%d-%c, Size: %dx%d, Format: %d\n",
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
version + 1, qr_ecc_level_names[ecc_level], size, size, format);
}
format_full = qr_annex_c1[(format << 2) + bitmask];
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
for (i = 1; i <= 8; i++) {
grid[(8 * size) + i] |= (format_full >> (15 - i)) & 0x01;
}
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
for (i = 7; i >= 1; i--) {
grid[(i * size) + 8] |= (format_full >> (i - 1)) & 0x01;
}
symbol->width = size;
symbol->rows = size;
for (i = 0; i < size; i++) {
for (j = 0; j < size; j++) {
if (grid[(i * size) + j] & 0x01) {
set_module(symbol, i, j);
}
}
symbol->row_height[i] = 1;
}
symbol->height = size;
return 0;
2016-02-17 23:52:48 +13:00
}
2017-05-21 10:37:50 +12:00
/* For UPNQR the symbol size and error correction capacity is fixed */
INTERNAL int upnqr(struct zint_symbol *symbol, unsigned char source[], int length) {
int i, j, r, est_binlen;
int ecc_level, version, target_codewords, blocks, size;
2017-05-21 20:20:03 +12:00
int bitmask, error_number;
int user_mask;
int size_squared;
struct zint_seg segs[1];
const int seg_count = 1;
const int fast_encode = symbol->input_mode & FAST_MODE;
const int debug_print = symbol->debug & ZINT_DEBUG_PRINT;
unsigned char *datastream;
unsigned char *fullstream;
unsigned char *grid;
unsigned int *ddata = (unsigned int *) z_alloca(sizeof(unsigned int) * length);
char *mode = (char *) z_alloca(length + 1);
unsigned char *preprocessed = (unsigned char *) z_alloca(length + 1);
2017-05-21 10:37:50 +12:00
symbol->eci = 4; /* Set before any processing */
user_mask = (symbol->option_3 >> 8) & 0x0F; /* User mask is pattern + 1, so >= 1 and <= 8 */
if (user_mask > 8) {
user_mask = 0; /* Ignore */
}
switch (symbol->input_mode & 0x07) {
2017-05-21 20:20:03 +12:00
case DATA_MODE:
2017-05-21 10:37:50 +12:00
/* Input is already in ISO-8859-2 format */
for (i = 0; i < length; i++) {
ddata[i] = source[i];
2017-05-21 10:37:50 +12:00
mode[i] = 'B';
}
break;
case GS1_MODE: /* Should never happen as checked before being called */
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxt(ZINT_ERROR_INVALID_OPTION, symbol, 571,
"UPNQR does not support GS1 data"); /* Not reached */
2017-05-21 20:20:03 +12:00
break;
case UNICODE_MODE:
error_number = utf8_to_eci(4, source, preprocessed, &length);
2017-05-21 20:20:03 +12:00
if (error_number != 0) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxt(error_number, symbol, 572, "Invalid character in input for ECI '4'");
2017-05-21 20:20:03 +12:00
}
for (i = 0; i < length; i++) {
ddata[i] = preprocessed[i];
2017-05-21 20:20:03 +12:00
mode[i] = 'B';
}
2017-05-21 10:37:50 +12:00
break;
}
segs[0].source = source;
segs[0].length = length;
segs[0].eci = 4;
est_binlen = qr_calc_binlen_segs(15, mode, ddata, segs, seg_count, NULL /*p_structapp*/, 1 /*mode_preset*/,
0 /*gs1*/, debug_print);
2017-10-24 08:37:52 +13:00
ecc_level = QR_LEVEL_M;
2017-05-21 10:37:50 +12:00
if (est_binlen > 3320) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxtf(ZINT_ERROR_TOO_LONG, symbol, 573, "Input too long, requires %d codewords (maximum 415)",
(est_binlen + 7) / 8);
2017-05-21 10:37:50 +12:00
}
version = 15; /* 77 x 77 */
2017-05-21 10:37:50 +12:00
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
target_codewords = qr_data_codewords[ecc_level][version - 1];
blocks = qr_blocks[ecc_level][version - 1];
datastream = (unsigned char *) z_alloca(target_codewords + 1);
fullstream = (unsigned char *) z_alloca(qr_total_codewords[version - 1] + 1);
2017-05-21 10:37:50 +12:00
(void) qr_binary_segs(datastream, version, target_codewords, mode, ddata, segs, seg_count, NULL /*p_structapp*/,
0 /*gs1*/, est_binlen, debug_print);
#ifdef ZINT_TEST
2019-12-17 06:31:52 +13:00
if (symbol->debug & ZINT_DEBUG_TEST) debug_test_codeword_dump(symbol, datastream, target_codewords);
#endif
qr_add_ecc(fullstream, datastream, version, target_codewords, blocks, debug_print);
2017-05-21 10:37:50 +12:00
size = qr_sizes[version - 1];
size_squared = size * size;
2017-05-21 10:37:50 +12:00
grid = (unsigned char *) z_alloca(size_squared);
memset(grid, 0, size_squared);
2017-05-21 10:37:50 +12:00
qr_setup_grid(grid, size, version);
qr_populate_grid(grid, size, size, fullstream, qr_total_codewords[version - 1]);
2017-10-24 08:37:52 +13:00
qr_add_version_info(grid, size, version);
2017-05-21 10:37:50 +12:00
bitmask = qr_apply_bitmask(grid, size, ecc_level, user_mask, fast_encode, debug_print);
2017-05-21 10:37:50 +12:00
qr_add_format_info(grid, size, ecc_level, bitmask);
2017-10-24 08:37:52 +13:00
2017-05-21 10:37:50 +12:00
symbol->width = size;
symbol->rows = size;
for (i = 0; i < size; i++) {
r = i * size;
2017-05-21 10:37:50 +12:00
for (j = 0; j < size; j++) {
if (grid[r + j] & 0x01) {
2017-05-21 10:37:50 +12:00
set_module(symbol, i, j);
}
}
symbol->row_height[i] = 1;
}
symbol->height = size;
2017-05-21 10:37:50 +12:00
return 0;
}
2017-10-24 08:34:31 +13:00
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
static const char rmqr_version_names[38][8] = {
"R7x43", "R7x59", "R7x77", "R7x99", "R7x139", "R9x43", "R9x59", "R9x77",
"R9x99", "R9x139", "R11x27", "R11x43", "R11x59", "R11x77", "R11x99", "R11x139",
"R13x27", "R13x43", "R13x59", "R13x77", "R13x99", "R13x139", "R15x43", "R15x59",
"R15x77", "R15x99", "R15x139", "R17x43", "R17x59", "R17x77", "R17x99", "R17x139",
"R7xW", "R9xW", "R11xW", "R13xW", "R15xW", "R17xW",
};
static void rmqr_setup_grid(unsigned char *grid, const int h_size, const int v_size) {
int i, j;
char alignment[] = {0x1F, 0x11, 0x15, 0x11, 0x1F};
int h_version, finder_position;
2017-10-24 08:37:52 +13:00
/* Add timing patterns - top and bottom */
for (i = 0; i < h_size; i++) {
if (i % 2) {
grid[i] = 0x20;
grid[((v_size - 1) * h_size) + i] = 0x20;
} else {
grid[i] = 0x21;
grid[((v_size - 1) * h_size) + i] = 0x21;
}
}
/* Add timing patterns - left and right */
for (i = 0; i < v_size; i++) {
if (i % 2) {
grid[i * h_size] = 0x20;
grid[(i * h_size) + (h_size - 1)] = 0x20;
} else {
grid[i * h_size] = 0x21;
grid[(i * h_size) + (h_size - 1)] = 0x21;
}
}
/* Add finder pattern */
qr_place_finder(grid, h_size, 0, 0); /* This works because finder is always top left */
/* Add finder sub-pattern to bottom right */
for (i = 0; i < 5; i++) {
for (j = 0; j < 5; j++) {
if (alignment[j] & 0x10 >> i) {
grid[((v_size - 5) * h_size) + (h_size * i) + (h_size - 5) + j] = 0x11;
} else {
grid[((v_size - 5) * h_size) + (h_size * i) + (h_size - 5) + j] = 0x10;
}
}
}
/* Add corner finder pattern - bottom left */
grid[(v_size - 2) * h_size] = 0x11;
grid[((v_size - 2) * h_size) + 1] = 0x10;
grid[((v_size - 1) * h_size) + 1] = 0x11;
/* Add corner finder pattern - top right */
grid[h_size - 2] = 0x11;
grid[(h_size * 2) - 2] = 0x10;
grid[(h_size * 2) - 1] = 0x11;
/* Add seperator */
for (i = 0; i < 7; i++) {
grid[(i * h_size) + 7] = 0x20;
}
if (v_size > 7) {
/* Note for v_size = 9 this overrides the bottom right corner finder pattern */
for (i = 0; i < 8; i++) {
grid[(7 * h_size) + i] = 0x20;
}
}
/* Add alignment patterns */
if (h_size > 27) {
h_version = 0; /* Suppress compiler warning [-Wmaybe-uninitialized] */
for (i = 0; i < 5; i++) {
if (h_size == rmqr_width[i]) {
h_version = i;
break;
}
}
for (i = 0; i < 4; i++) {
finder_position = rmqr_table_d1[(h_version * 4) + i];
if (finder_position != 0) {
for (j = 0; j < v_size; j++) {
if (j % 2) {
grid[(j * h_size) + finder_position] = 0x10;
} else {
grid[(j * h_size) + finder_position] = 0x11;
}
}
/* Top square */
grid[h_size + finder_position - 1] = 0x11;
grid[(h_size * 2) + finder_position - 1] = 0x11;
grid[h_size + finder_position + 1] = 0x11;
grid[(h_size * 2) + finder_position + 1] = 0x11;
/* Bottom square */
grid[(h_size * (v_size - 3)) + finder_position - 1] = 0x11;
grid[(h_size * (v_size - 2)) + finder_position - 1] = 0x11;
grid[(h_size * (v_size - 3)) + finder_position + 1] = 0x11;
grid[(h_size * (v_size - 2)) + finder_position + 1] = 0x11;
}
}
}
/* Reserve space for format information */
for (i = 0; i < 5; i++) {
for (j = 0; j < 3; j++) {
grid[(h_size * (i + 1)) + j + 8] = 0x20;
grid[(h_size * (v_size - 6)) + (h_size * i) + j + (h_size - 8)] = 0x20;
}
}
grid[(h_size * 1) + 11] = 0x20;
grid[(h_size * 2) + 11] = 0x20;
grid[(h_size * 3) + 11] = 0x20;
grid[(h_size * (v_size - 6)) + (h_size - 5)] = 0x20;
grid[(h_size * (v_size - 6)) + (h_size - 4)] = 0x20;
grid[(h_size * (v_size - 6)) + (h_size - 3)] = 0x20;
}
/* rMQR according to 2018 draft standard */
INTERNAL int rmqr(struct zint_symbol *symbol, struct zint_seg segs[], const int seg_count) {
int warn_number;
int i, j, est_binlen;
int ecc_level, autosize, version, max_cw, target_codewords, blocks, h_size, v_size;
int footprint, best_footprint, format_data;
unsigned int left_format_info, right_format_info;
const int gs1 = ((symbol->input_mode & 0x07) == GS1_MODE);
const int debug_print = symbol->debug & ZINT_DEBUG_PRINT;
const int eci_length_segs = get_eci_length_segs(segs, seg_count);
struct zint_seg *local_segs = (struct zint_seg *) z_alloca(sizeof(struct zint_seg) * seg_count);
unsigned int *ddata = (unsigned int *) z_alloca(sizeof(unsigned int) * eci_length_segs);
char *mode = (char *) z_alloca(eci_length_segs);
unsigned char *datastream;
unsigned char *fullstream;
unsigned char *grid;
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
if (symbol->option_1 == 1) {
return errtxt(ZINT_ERROR_INVALID_OPTION, symbol, 576, "Error correction level L not available in rMQR");
}
if (symbol->option_1 == 3) {
return errtxt(ZINT_ERROR_INVALID_OPTION, symbol, 577, "Error correction level Q not available in rMQR");
}
if ((symbol->option_2 < 0) || (symbol->option_2 > 38)) {
return errtxtf(ZINT_ERROR_INVALID_OPTION, symbol, 579, "Version '%d' out of range (1 to 38)",
symbol->option_2);
}
segs_cpy(symbol, segs, seg_count, local_segs);
warn_number = qr_prep_data(symbol, local_segs, seg_count, ddata);
if (warn_number >= ZINT_ERROR) {
return warn_number;
}
/* GS1 General Specifications 22.0 section 5.7.3 says ECIs not supported
for GS1 QR Code so check and return ZINT_WARN_NONCOMPLIANT if true */
if (gs1 && warn_number == 0) {
for (i = 0; i < seg_count; i++) {
if (local_segs[i].eci) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
warn_number = errtxt(ZINT_WARN_NONCOMPLIANT, symbol, 757,
"Using ECI in GS1 mode not supported by GS1 standards");
break;
}
}
}
est_binlen = qr_calc_binlen_segs(RMQR_VERSION + 31, mode, ddata, local_segs, seg_count, NULL /*p_structapp*/,
0 /*mode_preset*/, gs1, debug_print);
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
ecc_level = symbol->option_1 == 4 ? QR_LEVEL_H : QR_LEVEL_M;
max_cw = rmqr_data_codewords[ecc_level >> 1][31];
if (est_binlen > (8 * max_cw)) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxtf(ZINT_ERROR_TOO_LONG, symbol, 578,
"Input too long for ECC level %1$c, requires %2$d codewords (maximum %3$d)",
qr_ecc_level_names[ecc_level], (est_binlen + 7) / 8, max_cw);
}
version = 31; /* Set default to keep compiler happy */
if (symbol->option_2 == 0) {
/* Automatic symbol size */
autosize = 31;
best_footprint = rmqr_height[31] * rmqr_width[31];
for (version = 30; version >= 0; version--) {
est_binlen = qr_calc_binlen_segs(RMQR_VERSION + version, mode, ddata, local_segs, seg_count,
NULL /*p_structapp*/, 0 /*mode_preset*/, gs1, debug_print);
footprint = rmqr_height[version] * rmqr_width[version];
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
if (8 * rmqr_data_codewords[ecc_level >> 1][version] >= est_binlen) {
if (footprint < best_footprint) {
autosize = version;
best_footprint = footprint;
}
}
}
version = autosize;
est_binlen = qr_calc_binlen_segs(RMQR_VERSION + version, mode, ddata, local_segs, seg_count,
NULL /*p_structapp*/, 0 /*mode_preset*/, gs1, debug_print);
}
if ((symbol->option_2 >= 1) && (symbol->option_2 <= 32)) {
/* User specified symbol size */
version = symbol->option_2 - 1;
est_binlen = qr_calc_binlen_segs(RMQR_VERSION + version, mode, ddata, local_segs, seg_count,
NULL /*p_structapp*/, 0 /*mode_preset*/, gs1, debug_print);
}
if (symbol->option_2 >= 33) {
/* User has specified symbol height only */
version = rmqr_fixed_height_upper_bound[symbol->option_2 - 32];
for (i = version - 1; i > rmqr_fixed_height_upper_bound[symbol->option_2 - 33]; i--) {
est_binlen = qr_calc_binlen_segs(RMQR_VERSION + i, mode, ddata, local_segs, seg_count,
NULL /*p_structapp*/, 0 /*mode_preset*/, gs1, debug_print);
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
if (8 * rmqr_data_codewords[ecc_level >> 1][i] >= est_binlen) {
version = i;
}
}
est_binlen = qr_calc_binlen_segs(RMQR_VERSION + version, mode, ddata, local_segs, seg_count,
NULL /*p_structapp*/, 0 /*mode_preset*/, gs1, debug_print);
}
if (symbol->option_1 == -1) {
/* Detect if there is enough free space to increase ECC level */
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
if (est_binlen < rmqr_data_codewords[QR_LEVEL_H >> 1][version] * 8) {
ecc_level = QR_LEVEL_H;
}
}
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
target_codewords = rmqr_data_codewords[ecc_level >> 1][version];
blocks = rmqr_blocks[ecc_level >> 1][version];
if (est_binlen > (target_codewords * 8)) {
/* User has selected a symbol too small for the data */
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
assert(symbol->option_2 > 0);
return errtxtf(ZINT_ERROR_TOO_LONG, symbol, 560,
"Input too long for Version %1$d %2$s-%3$c, requires %4$d codewords (maximum %5$d)",
symbol->option_2, rmqr_version_names[symbol->option_2 - 1], qr_ecc_level_names[ecc_level],
(est_binlen + 7) / 8, target_codewords);
}
if (debug_print) {
printf("Minimum codewords: %d\n", (est_binlen + 7) / 8);
printf("Selected version: %d = R%dx%d-%c\n",
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
(version + 1), rmqr_height[version], rmqr_width[version], qr_ecc_level_names[ecc_level]);
printf("Number of data codewords in symbol: %d\n", target_codewords);
printf("Number of ECC blocks: %d\n", blocks);
}
datastream = (unsigned char *) z_alloca(target_codewords + 1);
fullstream = (unsigned char *) z_alloca(rmqr_total_codewords[version] + 1);
(void) qr_binary_segs(datastream, RMQR_VERSION + version, target_codewords, mode, ddata, local_segs, seg_count,
NULL /*p_structapp*/, gs1, est_binlen, debug_print);
#ifdef ZINT_TEST
2019-12-17 06:31:52 +13:00
if (symbol->debug & ZINT_DEBUG_TEST) debug_test_codeword_dump(symbol, datastream, target_codewords);
#endif
qr_add_ecc(fullstream, datastream, RMQR_VERSION + version, target_codewords, blocks, debug_print);
h_size = rmqr_width[version];
v_size = rmqr_height[version];
grid = (unsigned char *) z_alloca(h_size * v_size);
memset(grid, 0, h_size * v_size);
rmqr_setup_grid(grid, h_size, v_size);
qr_populate_grid(grid, h_size, v_size, fullstream, rmqr_total_codewords[version]);
/* apply bitmask */
for (i = 0; i < v_size; i++) {
int r = i * h_size;
for (j = 0; j < h_size; j++) {
if ((grid[r + j] & 0xf0) == 0) {
/* This is a data module */
if (((i / 2) + (j / 3)) % 2 == 0) { /* < This is the data mask from section 7.8.2 */
/* This module needs to be changed */
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
grid[r + j] ^= 0x01;
}
}
}
}
/* add format information */
format_data = version;
if (ecc_level == QR_LEVEL_H) {
format_data += 32;
}
left_format_info = rmqr_format_info_left[format_data];
right_format_info = rmqr_format_info_right[format_data];
for (i = 0; i < 5; i++) {
for (j = 0; j < 3; j++) {
grid[(h_size * (i + 1)) + j + 8] = (left_format_info >> ((j * 5) + i)) & 0x01;
grid[(h_size * (v_size - 6)) + (h_size * i) + j + (h_size - 8)]
= (right_format_info >> ((j * 5) + i)) & 0x01;
}
}
grid[(h_size * 1) + 11] = (left_format_info >> 15) & 0x01;
grid[(h_size * 2) + 11] = (left_format_info >> 16) & 0x01;
grid[(h_size * 3) + 11] = (left_format_info >> 17) & 0x01;
grid[(h_size * (v_size - 6)) + (h_size - 5)] = (right_format_info >> 15) & 0x01;
grid[(h_size * (v_size - 6)) + (h_size - 4)] = (right_format_info >> 16) & 0x01;
grid[(h_size * (v_size - 6)) + (h_size - 3)] = (right_format_info >> 17) & 0x01;
symbol->width = h_size;
symbol->rows = v_size;
for (i = 0; i < v_size; i++) {
int r = i * h_size;
for (j = 0; j < h_size; j++) {
if (grid[r + j] & 0x01) {
set_module(symbol, i, j);
}
}
symbol->row_height[i] = 1;
}
symbol->height = v_size;
return warn_number;
}
/* vim: set ts=4 sw=4 et : */