zint/backend/tests/test_postal.c

657 lines
40 KiB
C
Raw Normal View History

/*
libzint - the open source barcode library
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
Copyright (C) 2019-2024 Robin Stuart <rstuart114@gmail.com>
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.
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.
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.
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
SUCH DAMAGE.
*/
/* SPDX-License-Identifier: BSD-3-Clause */
/* USPS Publication 25 (July 2003) Designing Letter and Reply Mail https://web.archive.org/web/20050118015758/http://www.siemons.com/forms/pdf/designing_letter_reply_mail.pdf */
/* USPS DMM Domestic Mail Manual https://pe.usps.com/DMM300 */
/* USPS Publication 197 (Sept 2004) Confirm User Guide https://web.archive.org/web/20060505214851/https://mailtracking.usps.com/mtr/resources/documents/Guide.pdf */
#include "testcommon.h"
static void test_large(const testCtx *const p_ctx) {
int debug = p_ctx->debug;
struct item {
int symbology;
char *pattern;
int length;
int ret;
int expected_rows;
int expected_width;
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
char *expected_errtxt;
};
/* s/\/\*[ 0-9]*\*\//\=printf("\/\*%3d*\/", line(".") - line("'<")): */
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 struct item data[] = {
/* 0*/ { BARCODE_FLAT, "1", 128, 0, 1, 1152, "" },
/* 1*/ { BARCODE_FLAT, "1", 129, ZINT_ERROR_TOO_LONG, -1, -1, "Error 494: Input length 129 too long (maximum 128)" },
/* 2*/ { BARCODE_POSTNET, "1", 11, 0, 2, 123, "" },
/* 3*/ { BARCODE_POSTNET, "1", 12, ZINT_WARN_NONCOMPLIANT, 2, 133, "Warning 479: Input length 12 is not standard (5, 9 or 11)" },
/* 4*/ { BARCODE_POSTNET, "1", 38, ZINT_WARN_NONCOMPLIANT, 2, 393, "Warning 479: Input length 38 is not standard (5, 9 or 11)" },
/* 5*/ { BARCODE_POSTNET, "1", 39, ZINT_ERROR_TOO_LONG, -1, -1, "Error 480: Input length 39 too long (maximum 38)" },
/* 6*/ { BARCODE_FIM, "D", 1, 0, 1, 17, "" },
/* 7*/ { BARCODE_FIM, "D", 2, ZINT_ERROR_TOO_LONG, -1, -1, "Error 486: Input length 2 too long (maximum 1)" },
/* 8*/ { BARCODE_CEPNET, "1", 8, 0, 2, 93, "" },
/* 9*/ { BARCODE_CEPNET, "1", 7, ZINT_WARN_NONCOMPLIANT, 2, 83, "Warning 780: Input length 7 wrong (should be 8 digits)" },
/* 10*/ { BARCODE_CEPNET, "1", 9, ZINT_WARN_NONCOMPLIANT, 2, 103, "Warning 780: Input length 9 wrong (should be 8 digits)" },
/* 11*/ { BARCODE_CEPNET, "1", 39, ZINT_ERROR_TOO_LONG, -1, -1, "Error 480: Input length 39 too long (maximum 38)" },
/* 12*/ { BARCODE_RM4SCC, "1", 50, 0, 3, 411, "" },
/* 13*/ { BARCODE_RM4SCC, "1", 51, ZINT_ERROR_TOO_LONG, -1, -1, "Error 488: Input length 51 too long (maximum 50)" },
/* 14*/ { BARCODE_JAPANPOST, "1", 20, 0, 3, 133, "" },
/* 15*/ { BARCODE_JAPANPOST, "1", 21, ZINT_ERROR_TOO_LONG, -1, -1, "Error 496: Input length 21 too long (maximum 20)" },
/* 16*/ { BARCODE_JAPANPOST, "A", 10, 0, 3, 133, "" },
/* 17*/ { BARCODE_JAPANPOST, "A", 11, ZINT_ERROR_TOO_LONG, -1, -1, "Error 477: Input too long, requires too many symbol characters (maximum 20)" },
/* 18*/ { BARCODE_KOREAPOST, "1", 6, 0, 1, 162, "" },
/* 19*/ { BARCODE_KOREAPOST, "1", 7, ZINT_ERROR_TOO_LONG, -1, -1, "Error 484: Input length 7 too long (maximum 6)" },
/* 20*/ { BARCODE_PLANET, "1", 13, 0, 2, 143, "" },
/* 21*/ { BARCODE_PLANET, "1", 14, ZINT_WARN_NONCOMPLIANT, 2, 153, "Warning 478: Input length 14 is not standard (11 or 13)" },
/* 22*/ { BARCODE_PLANET, "1", 38, ZINT_WARN_NONCOMPLIANT, 2, 393, "Warning 478: Input length 38 is not standard (11 or 13)" },
/* 23*/ { BARCODE_PLANET, "1", 39, ZINT_ERROR_TOO_LONG, -1, -1, "Error 482: Input length 39 too long (maximum 38)" },
/* 24*/ { BARCODE_KIX, "1", 18, 0, 3, 143, "" },
/* 25*/ { BARCODE_KIX, "1", 19, ZINT_ERROR_TOO_LONG, -1, -1, "Error 490: Input length 19 too long (maximum 18)" },
/* 26*/ { BARCODE_DAFT, "D", 576, 0, 3, 1151, "" },
/* 27*/ { BARCODE_DAFT, "D", 577, ZINT_ERROR_TOO_LONG, -1, -1, "Error 492: Input length 577 too long (maximum 576)" },
};
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 data_size = ARRAY_SIZE(data);
int i, length, ret;
struct zint_symbol *symbol = NULL;
char data_buf[4096];
testStartSymbol("test_large", &symbol);
for (i = 0; i < data_size; i++) {
if (testContinue(p_ctx, i)) continue;
symbol = ZBarcode_Create();
assert_nonnull(symbol, "Symbol not created\n");
testUtilStrCpyRepeat(data_buf, data[i].pattern, data[i].length);
assert_equal(data[i].length, (int) strlen(data_buf), "i:%d length %d != strlen(data_buf) %d\n", i, data[i].length, (int) strlen(data_buf));
length = testUtilSetSymbol(symbol, data[i].symbology, -1 /*input_mode*/, -1 /*eci*/, -1 /*option_1*/, -1, -1, -1 /*output_options*/, data_buf, data[i].length, debug);
ret = ZBarcode_Encode(symbol, (unsigned char *) data_buf, length);
assert_equal(ret, data[i].ret, "i:%d ZBarcode_Encode ret %d != %d (%s)\n", i, ret, data[i].ret, symbol->errtxt);
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_zero(strcmp(symbol->errtxt, data[i].expected_errtxt), "i:%d strcmp(%s, %s) != 0\n", i, symbol->errtxt, data[i].expected_errtxt);
if (ret < ZINT_ERROR) {
assert_equal(symbol->rows, data[i].expected_rows, "i:%d symbol->rows %d != %d\n", i, symbol->rows, data[i].expected_rows);
assert_equal(symbol->width, data[i].expected_width, "i:%d symbol->width %d != %d\n", i, symbol->width, data[i].expected_width);
}
ZBarcode_Delete(symbol);
}
testFinish();
}
static void test_koreapost(const testCtx *const p_ctx) {
int debug = p_ctx->debug;
struct item {
char *data;
int ret_encode;
int ret_vector;
float expected_height;
int expected_rows;
int expected_width;
};
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 struct item data[] = {
/* 0*/ { "123456", 0, 0, 50, 1, 167 },
};
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 data_size = ARRAY_SIZE(data);
int i, length, ret;
struct zint_symbol *symbol = NULL;
testStartSymbol("test_koreapost", &symbol);
for (i = 0; i < data_size; i++) {
if (testContinue(p_ctx, i)) continue;
symbol = ZBarcode_Create();
assert_nonnull(symbol, "Symbol not created\n");
length = testUtilSetSymbol(symbol, BARCODE_KOREAPOST, -1 /*input_mode*/, -1 /*eci*/, -1 /*option_1*/, -1, -1, -1 /*output_options*/, data[i].data, -1, debug);
ret = ZBarcode_Encode(symbol, (unsigned char *) data[i].data, length);
assert_equal(ret, data[i].ret_encode, "i:%d ZBarcode_Encode ret %d != %d\n", i, ret, data[i].ret_encode);
if (ret < ZINT_ERROR) {
assert_equal(symbol->height, data[i].expected_height, "i:%d symbol->height %.8g != %.8g\n", i, symbol->height, data[i].expected_height);
assert_equal(symbol->rows, data[i].expected_rows, "i:%d symbol->rows %d != %d\n", i, symbol->rows, data[i].expected_rows);
assert_equal(symbol->width, data[i].expected_width, "i:%d symbol->width %d != %d\n", i, symbol->width, data[i].expected_width);
ret = ZBarcode_Buffer_Vector(symbol, 0);
assert_equal(ret, data[i].ret_vector, "i:%d ZBarcode_Buffer_Vector ret %d != %d\n", i, ret, data[i].ret_vector);
}
ZBarcode_Delete(symbol);
}
testFinish();
}
static void test_japanpost(const testCtx *const p_ctx) {
int debug = p_ctx->debug;
struct item {
char *data;
int ret_encode;
int ret_vector;
float expected_height;
int expected_rows;
int expected_width;
char *comment;
};
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 struct item data[] = {
/* 0*/ { "123", 0, 0, 8, 3, 133, "Check 3" },
/* 1*/ { "123456-AB", 0, 0, 8, 3, 133, "Check 10" },
/* 2*/ { "123456", 0, 0, 8, 3, 133, "Check 11" },
/* 3*/ { "999980-KZ", 0, 0, 8, 3, 133, "Check 18" },
/* 4*/ { "987654-TU", 0, 0, 8, 3, 133, "Check 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
const int data_size = ARRAY_SIZE(data);
int i, length, ret;
struct zint_symbol *symbol = NULL;
testStartSymbol("test_japanpost", &symbol);
for (i = 0; i < data_size; i++) {
if (testContinue(p_ctx, i)) continue;
symbol = ZBarcode_Create();
assert_nonnull(symbol, "Symbol not created\n");
length = testUtilSetSymbol(symbol, BARCODE_JAPANPOST, -1 /*input_mode*/, -1 /*eci*/, -1 /*option_1*/, -1, -1, -1 /*output_options*/, data[i].data, -1, debug);
ret = ZBarcode_Encode(symbol, (unsigned char *) data[i].data, length);
assert_equal(ret, data[i].ret_encode, "i:%d ZBarcode_Encode ret %d != %d\n", i, ret, data[i].ret_encode);
if (ret < ZINT_ERROR) {
assert_equal(symbol->height, data[i].expected_height, "i:%d symbol->height %.8g != %.8g\n", i, symbol->height, data[i].expected_height);
assert_equal(symbol->rows, data[i].expected_rows, "i:%d symbol->rows %d != %d\n", i, symbol->rows, data[i].expected_rows);
assert_equal(symbol->width, data[i].expected_width, "i:%d symbol->width %d != %d\n", i, symbol->width, data[i].expected_width);
ret = ZBarcode_Buffer_Vector(symbol, 0);
assert_equal(ret, data[i].ret_vector, "i:%d ZBarcode_Buffer_Vector ret %d != %d\n", i, ret, data[i].ret_vector);
}
ZBarcode_Delete(symbol);
}
testFinish();
}
static void test_input(const testCtx *const p_ctx) {
int debug = p_ctx->debug;
struct item {
int symbology;
int option_2;
float height;
char *data;
int ret;
int expected_rows;
int expected_width;
float expected_height;
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
char *expected_errtxt;
int bwipp_cmp;
char *comment;
};
/* s/\/\*[ 0-9]*\*\//\=printf("\/\*%3d*\/", line(".") - line("'<")): */
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 struct item data[] = {
/* 0*/ { BARCODE_FLAT, -1, 0, "1234567890", 0, 1, 90, 50, "", 1, "" },
/* 1*/ { BARCODE_FLAT, -1, 0, "A", ZINT_ERROR_INVALID_DATA, -1, -1, -1, "Error 495: Invalid character at position 1 in input (digits only)", 1, "" },
/* 2*/ { BARCODE_POSTNET, -1, 0, "12345", 0, 2, 63, 12, "", 1, "" },
/* 3*/ { BARCODE_POSTNET, -1, 0, "123457689", 0, 2, 103, 12, "", 1, "" },
/* 4*/ { BARCODE_POSTNET, -1, 0, "12345768901", 0, 2, 123, 12, "", 1, "" },
/* 5*/ { BARCODE_POSTNET, -1, 0, "0", ZINT_WARN_NONCOMPLIANT, 2, 23, 12, "Warning 479: Input length 1 is not standard (5, 9 or 11)", 0, "BWIPP requires standard lengths" },
/* 6*/ { BARCODE_POSTNET, -1, 0, "1234", ZINT_WARN_NONCOMPLIANT, 2, 53, 12, "Warning 479: Input length 4 is not standard (5, 9 or 11)", 0, "BWIPP requires standard lengths" },
/* 7*/ { BARCODE_POSTNET, -1, 0, "123456", ZINT_WARN_NONCOMPLIANT, 2, 73, 12, "Warning 479: Input length 6 is not standard (5, 9 or 11)", 0, "BWIPP requires standard lengths" },
/* 8*/ { BARCODE_POSTNET, -1, 0, "123456789012", ZINT_WARN_NONCOMPLIANT, 2, 133, 12, "Warning 479: Input length 12 is not standard (5, 9 or 11)", 0, "BWIPP requires standard lengths" },
/* 9*/ { BARCODE_POSTNET, -1, 0, "1234A", ZINT_ERROR_INVALID_DATA, -1, -1, -1, "Error 481: Invalid character at position 5 in input (digits only)", 1, "" },
/* 10*/ { BARCODE_POSTNET, -1, 0.9, "12345", 0, 2, 63, 1, "", 1, "" },
/* 11*/ { BARCODE_FIM, -1, 0, "a", 0, 1, 17, 50, "", 1, "" },
/* 12*/ { BARCODE_FIM, -1, 0, "b", 0, 1, 17, 50, "", 1, "" },
/* 13*/ { BARCODE_FIM, -1, 0, "c", 0, 1, 17, 50, "", 1, "" },
/* 14*/ { BARCODE_FIM, -1, 0, "d", 0, 1, 17, 50, "", 1, "" },
/* 15*/ { BARCODE_FIM, -1, 0, "ad", ZINT_ERROR_TOO_LONG, -1, -1, -1, "Error 486: Input length 2 too long (maximum 1)", 1, "" },
/* 16*/ { BARCODE_FIM, -1, 0, "e", 0, 1, 17, 50, "", 1, "" },
/* 17*/ { BARCODE_FIM, -1, 0, "f", ZINT_ERROR_INVALID_DATA, -1, -1, -1, "Error 487: Invalid character in input (\"A\", \"B\", \"C\", \"D\" or \"E\" only)", 1, "" },
/* 18*/ { BARCODE_CEPNET, -1, 0, "1234567A", ZINT_ERROR_INVALID_DATA, -1, -1, -1, "Error 481: Invalid character at position 8 in input (digits only)", 1, "" },
/* 19*/ { BARCODE_RM4SCC, -1, 0, "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ", 0, 3, 299, 8, "", 1, "" },
/* 20*/ { BARCODE_RM4SCC, -1, 0, "a", 0, 3, 19, 8, "", 1, "" }, /* Converts to upper */
/* 21*/ { BARCODE_RM4SCC, -1, 0, ",", ZINT_ERROR_INVALID_DATA, -1, -1, -1, "Error 489: Invalid character at position 1 in input (alphanumerics only)", 1, "" },
/* 22*/ { BARCODE_JAPANPOST, -1, 0, "1234567890-ABCD", 0, 3, 133, 8, "", 1, "" }, /* 19 symbol chars */
/* 23*/ { BARCODE_JAPANPOST, -1, 0, "1234567890-ABCD1", 0, 3, 133, 8, "", 1, "" }, /* 20 symbol chars */
/* 24*/ { BARCODE_JAPANPOST, -1, 0, "1234567890-ABCDE", ZINT_ERROR_TOO_LONG, -1, -1, -1, "Error 477: Input too long, requires too many symbol characters (maximum 20)", 1, "" }, /* 21 symbol chars */
/* 25*/ { BARCODE_JAPANPOST, -1, 0, "1234567890-ABCD12", ZINT_ERROR_TOO_LONG, -1, -1, -1, "Error 477: Input too long, requires too many symbol characters (maximum 20)", 1, "" }, /* 21 symbol chars */
/* 26*/ { BARCODE_JAPANPOST, -1, 0, "1234567890ABCDE", 0, 3, 133, 8, "", 1, "" }, /* 20 symbol chars */
/* 27*/ { BARCODE_JAPANPOST, -1, 0, "a", 0, 3, 133, 8, "", 1, "" }, /* Converts to upper */
/* 28*/ { BARCODE_JAPANPOST, -1, 0, ",", ZINT_ERROR_INVALID_DATA, -1, -1, -1, "Error 497: Invalid character at position 1 in input (alphanumerics and \"-\" only)", 1, "" },
/* 29*/ { BARCODE_KOREAPOST, -1, 0, "123456", 0, 1, 167, 50, "", 1, "" },
/* 30*/ { BARCODE_KOREAPOST, -1, 0, "A", ZINT_ERROR_INVALID_DATA, -1, -1, -1, "Error 485: Invalid character at position 1 in input (digits only)", 1, "" },
/* 31*/ { BARCODE_PLANET, -1, 0, "12345678901", 0, 2, 123, 12, "", 1, "" },
/* 32*/ { BARCODE_PLANET, -1, 0, "1234567890123", 0, 2, 143, 12, "", 1, "" },
/* 33*/ { BARCODE_PLANET, -1, 0, "0", ZINT_WARN_NONCOMPLIANT, 2, 23, 12, "Warning 478: Input length 1 is not standard (11 or 13)", 0, "BWIPP requires standard lengths" },
/* 34*/ { BARCODE_PLANET, -1, 0, "1234567890", ZINT_WARN_NONCOMPLIANT, 2, 113, 12, "Warning 478: Input length 10 is not standard (11 or 13)", 0, "BWIPP requires standard lengths" },
/* 35*/ { BARCODE_PLANET, -1, 0, "123456789012", ZINT_WARN_NONCOMPLIANT, 2, 133, 12, "Warning 478: Input length 12 is not standard (11 or 13)", 0, "BWIPP requires standard lengths" },
/* 36*/ { BARCODE_PLANET, -1, 0, "12345678901234", ZINT_WARN_NONCOMPLIANT, 2, 153, 12, "Warning 478: Input length 14 is not standard (11 or 13)", 0, "BWIPP requires standard lengths" },
/* 37*/ { BARCODE_PLANET, -1, 0, "1234567890A", ZINT_ERROR_INVALID_DATA, -1, -1, -1, "Error 483: Invalid character at position 11 in input (digits only)", 1, "" },
/* 38*/ { BARCODE_KIX, -1, 0, "0123456789ABCDEFGH", 0, 3, 143, 8, "", 1, "" },
/* 39*/ { BARCODE_KIX, -1, 0, "a", 0, 3, 7, 8, "", 1, "" }, /* Converts to upper */
/* 40*/ { BARCODE_KIX, -1, 0, ",", ZINT_ERROR_INVALID_DATA, -1, -1, -1, "Error 491: Invalid character at position 1 in input (alphanumerics only)", 1, "" },
/* 41*/ { BARCODE_DAFT, -1, 0, "DAFT", 0, 3, 7, 8, "", 1, "" },
/* 42*/ { BARCODE_DAFT, -1, 0, "a", 0, 3, 1, 8, "", 1, "" }, /* Converts to upper */
/* 43*/ { BARCODE_DAFT, -1, 0, "B", ZINT_ERROR_INVALID_DATA, -1, -1, -1, "Error 493: Invalid character at position 1 in input (\"D\", \"A\", \"F\" and \"T\" only)", 1, "" },
/* 44*/ { BARCODE_DAFT, -1, 1.9, "DAFT", 0, 3, 7, 2, "", 1, "" },
/* 45*/ { BARCODE_DAFT, 500, 0.9, "DAFT", 0, 3, 7, 2, "", 1, "" }, /* 50% ratio */
/* 46*/ { BARCODE_DAFT, 500, 0.4, "DAFT", 0, 3, 7, 8, "", 1, "" }, /* 50% ratio */
};
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 data_size = ARRAY_SIZE(data);
int i, length, ret;
struct zint_symbol *symbol = NULL;
char cmp_buf[8192];
char cmp_msg[1024];
int do_bwipp = (debug & ZINT_DEBUG_TEST_BWIPP) && testUtilHaveGhostscript(); /* Only do BWIPP test if asked, too slow otherwise */
testStartSymbol("test_input", &symbol);
for (i = 0; i < data_size; i++) {
if (testContinue(p_ctx, i)) continue;
symbol = ZBarcode_Create();
assert_nonnull(symbol, "Symbol not created\n");
length = testUtilSetSymbol(symbol, data[i].symbology, -1 /*input_mode*/, -1 /*eci*/, -1 /*option_1*/, data[i].option_2, -1, -1 /*output_options*/, data[i].data, -1, debug);
if (data[i].height) {
symbol->height = data[i].height;
}
ret = ZBarcode_Encode(symbol, (unsigned char *) data[i].data, length);
assert_equal(ret, data[i].ret, "i:%d ZBarcode_Encode ret %d != %d (%s)\n", i, ret, data[i].ret, symbol->errtxt);
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_zero(strcmp(symbol->errtxt, data[i].expected_errtxt), "i:%d strcmp(%s, %s) != 0\n", i, symbol->errtxt, data[i].expected_errtxt);
if (ret < ZINT_ERROR) {
assert_equal(symbol->rows, data[i].expected_rows, "i:%d symbol->rows %d != %d\n", i, symbol->rows, data[i].expected_rows);
assert_equal(symbol->width, data[i].expected_width, "i:%d symbol->width %d != %d\n", i, symbol->width, data[i].expected_width);
assert_equal(symbol->height, data[i].expected_height, "i:%d symbol->height %g != %g\n", i, symbol->height, data[i].expected_height);
if (do_bwipp && testUtilCanBwipp(i, symbol, -1, -1, -1, debug)) {
if (!data[i].bwipp_cmp) {
if (debug & ZINT_DEBUG_TEST_PRINT) printf("i:%d %s not BWIPP compatible (%s)\n", i, testUtilBarcodeName(symbol->symbology), data[i].comment);
} else {
char modules_dump[4096];
assert_notequal(testUtilModulesDump(symbol, modules_dump, sizeof(modules_dump)), -1, "i:%d testUtilModulesDump == -1\n", i);
ret = testUtilBwipp(i, symbol, -1, -1, -1, data[i].data, length, NULL, cmp_buf, sizeof(cmp_buf), NULL);
assert_zero(ret, "i:%d %s testUtilBwipp ret %d != 0\n", i, testUtilBarcodeName(symbol->symbology), ret);
ret = testUtilBwippCmp(symbol, cmp_msg, cmp_buf, modules_dump);
assert_zero(ret, "i:%d %s testUtilBwippCmp %d != 0 %s\n actual: %s\nexpected: %s\n",
i, testUtilBarcodeName(symbol->symbology), ret, cmp_msg, cmp_buf, modules_dump);
}
}
}
ZBarcode_Delete(symbol);
}
testFinish();
}
static void test_encode(const testCtx *const p_ctx) {
int debug = p_ctx->debug;
struct item {
int symbology;
char *data;
int ret;
int expected_rows;
int expected_width;
char *comment;
char *expected;
};
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 struct item data[] = {
/* 0*/ { BARCODE_FLAT, "1304056", 0, 1, 63, "Verified manually against tec-it",
"100000000001000000000000000000100000000000000000010000000001000"
},
/* 1*/ { BARCODE_POSTNET, "12345678901", 0, 2, 123, "USPS Publication 25 (2003) Exhibit 4-1",
"100000001010000010001000001010000010000010001000100000101000001000000010100000100010001000001010000000000000101000100000101"
"101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
},
/* 2*/ { BARCODE_POSTNET, "555551237", 0, 2, 103, "Verified manually against tec-it",
"1000100010000010001000001000100000100010000010001000000000101000001000100000101000100000001000001000101"
"1010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
},
/* 3*/ { BARCODE_FIM, "C", 0, 1, 17, "USPS DMM Exhibit 8.2.0 FIM C",
"10100010001000101"
},
/* 4*/ { BARCODE_FIM, "E", 0, 1, 17, "",
"10001000000010001"
},
/* 5*/ { BARCODE_CEPNET, "12345678", 0, 2, 93, "CEPNet e Código Bidimensional Datamatrix 2D (26/05/2021) Figura 8",
"100000001010000010001000001010000010000010001000100000101000001000000010100000100000100000101"
"101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
},
/* 6*/ { BARCODE_CEPNET, "36400000", 0, 2, 93, "CEPNet e Código Bidimensional Datamatrix 2D (26/05/2021) Figura 10",
"100000101000001010000000100000101010000000101000000010100000001010000000101000000010000000101"
"101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
},
/* 7*/ { BARCODE_RM4SCC, "BX11LT1A", 0, 3, 75, "Verified manually against tec-it",
"100010001010100000000010100000101010000010100010000000101000100010100000101"
"101010101010101010101010101010101010101010101010101010101010101010101010101"
"001010000010000010001000100010001010000010101000000010001010001000000010101"
},
/* 8*/ { BARCODE_RM4SCC, "W1J0TR01", 0, 3, 75, "Verified manually against tec-it",
"101010000000001010100000100000101010001000100010000000101000001010101000001"
"101010101010101010101010101010101010101010101010101010101010101010101010101"
"000010100000100010001000100000101010100000100000100000101000100010001010001"
},
/* 9*/ { BARCODE_RM4SCC, "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ", 0, 3, 299, "Verified manually against tec-it",
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
"10000010100000101000001010000010100000101000001010001000100010001000100010001000100010001000100010001010000010100000101000001010000010100000101000100000101000001010000010100000101000001010000010100010001000100010001000100010001000100010001000101000001010000010100000101000001010000010100000101000001"
"10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
"00000010100010001000101000100000101000100010100000000010100010001000101000100000101000100010100000000010100010001000101000100000101000100010100000000010100010001000101000100000101000100010100000000010100010001000101000100000101000100010100000000010100010001000101000100000101000100010100000101000001"
},
/* 10*/ { BARCODE_JAPANPOST, "15400233-16-4-205", 0, 3, 133, "Zip/Barcode Manual p.6 1st example; verified manually against tec-it",
"1000101000100010101000100000100000100010001010001010001000101000001010001000101000001000100010100000100010000010000010000010001010001"
"1010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
"1010101000100010100010100000100000101000101000101000001000101000100010001000100010001000101000100000100010001000001000001000100010101"
},
/* 11*/ { BARCODE_JAPANPOST, "350110622-1A308", 0, 3, 133, "Zip/Barcode Manual p.6 2nd example; verified manually against tec-it",
"1000001010100010100000101000101000100000001010100010100010001000101000001000100000001010100000100010000010000010000010000010100010001"
"1010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
"1010101000100010100000101000101000100000100010101000101000001000101000100000100000101000100000001010001000001000001000001000100010101"
},
/* 12*/ { BARCODE_JAPANPOST, "12345671-2-3", 0, 3, 133, "Verified manually against tec-it",
"1000101000100010001010101000100010001010101000101000001000100010001000001010000010000010000010000010000010000010000010000010100010001"
"1010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
"1010101000101000101000100010100010100010001010101000001000101000001000101000001000001000001000001000001000001000001000001000100010101"
},
/* 13*/ { BARCODE_JAPANPOST, "1234567BCDEFG", 0, 3, 133, "Verified manually against tec-it",
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
"1000101000100010001010101000100010001010101000001000101000001000100010001000001010001000101000001000100010001000001010000010101000001"
"1010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
"1010101000101000101000100010100010100010001010100000101000100000101000100000101000100000100010100000100010100000100010001000100010101"
},
/* 14*/ { BARCODE_JAPANPOST, "8901234HIJKLM", 0, 3, 133, "Verified manually against tec-it",
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
"1000100010001010100000101000100010001010101000001000101000001000100010001000001010000010100000000010101000000010100010000010100000001"
"1010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
"1010001010001010100000101000101000101000100010100000001010100000001010100000001010100000100000100000101000100000101000001000000010101"
},
/* 15*/ { BARCODE_JAPANPOST, "0987654NOPQRS", 0, 3, 133, "Verified manually against tec-it",
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
"1000100000001010100010101000001010100010101000000010001010000010101000000010100010000010001010000010101000000010100010000010100000001"
"1010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
"1010100000001010001010001010100010100010100010100000101000100000100010100000100010100000100010100000001010100000001010001000001000101"
},
/* 16*/ { BARCODE_JAPANPOST, "3210987TUVWXY", 0, 3, 133, "Verified manually against tec-it",
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
"1000001010100010101000100000001010100010101000000010001010100000100000100000101000100000100010100000001010100000101000000010000010001"
"1010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
"1010101000101000101000100000001010001010001010100000001010001000100000001000101000001000101000001000101000001000100010001000100000101"
},
/* 17*/ { BARCODE_KOREAPOST, "010230", 0, 1, 167, "Verified manually against tec-it",
"10001000100000000000100010000000000010001000100000001000000010001000100010001000100000000000100000000001000100010001000100010001000000000001000000010001000000010001000"
},
/* 18*/ { BARCODE_KOREAPOST, "923457", 0, 1, 168, "Verified manually against tec-it",
"000010001000100000001000100000001000000010001000000010001000000010001000100000000000100010001000000010000000100010001000100010000000100000001000100010001000000000001000"
},
/* 19*/ { BARCODE_PLANET, "4012345235636", 0, 2, 143, "USPS Publication 197 (2004) Exhibit 4; verified manually against tec-it",
"10100010100000001010101010100000101000100010100000101000101000100010001010100010001010000010100010001010000010101010000010100000101010000010101"
"10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
},
/* 20*/ { BARCODE_PLANET, "40123452356", 0, 2, 123, "Verified manually against tec-it",
"101000101000000010101010101000001010001000101000001010001010001000100010101000100010100000101000100010100000101010001000101"
"101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
},
/* 21*/ { BARCODE_PLANET, "5020140235635", 0, 2, 143, "USPS Publication 197 (2004) Exhibit 6; verified manually against tec-it",
"10100010001000001010101010001000000010101010101000001000101000000010101010100010001010000010100010001010000010101010000010100010001010001010001"
"10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
},
/* 22*/ { BARCODE_KIX, "2500GG30250", 0, 3, 87, "PostNL Handleiding KIX code Section 2.1 Example 1",
"000010100000101000001010000010100010100000101000000010100000101000001010000010100000101"
"101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
"001010001010000000001010000010101000100010001000100000100000101000101000101000000000101"
},
/* 23*/ { BARCODE_KIX, "2130VA80430", 0, 3, 87, "PostNL Handleiding KIX code Section 2.1 Example 2",
"000010100000101000001010000010101010000000100010001000100000101000001010000010100000101"
"101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
"001010000010001010000010000010100010001010001000001010000000101010001000100000100000101"
},
/* 24*/ { BARCODE_KIX, "1231GF156X2", 0, 3, 87, "PostNL Handleiding KIX code Section 2.1 Example 3",
"000010100000101000001010000010100010100000101000000010100000101000100010101000000000101"
"101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
"001000100010100010000010001000101000100010000010001000101010000000001010100000100010100"
},
/* 25*/ { BARCODE_KIX, "1231FZ13Xhs", 0, 3, 87, "PostNL Handleiding KIX code Section 2.1 Example 4",
"000010100000101000001010000010100010100010100000000010100000101010100000001010001000100"
"101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
"001000100010100010000010001000101000001010100000001000101000001010000010101000001000100"
},
/* 26*/ { BARCODE_KIX, "1234567890ABCDEFGH", 0, 3, 143, "Verified manually against tec-it",
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
"00001010000010100000101000001010000010100010001000100010001000100010001000001010001000100010001000101000001010000010100000101000001010000010100"
"10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
"00100010001010001000001010001000101000000000101000100010001010001000001000001010100010001010000000001010001000100010100010000010100010001010000"
},
/* 27*/ { BARCODE_KIX, "IJKLMNOPQRSTUVWXYZ", 0, 3, 143, "Verified manually against tec-it",
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
"10000010100000101000001010000010100000101000001010001000100010001000100010001000100010001000100010100000101000001010000010100000101000001010000"
"10101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101"
"00001010001000100010100010000010100010001010000000001010001000100010100010000010100010001010000000001010001000100010100010000010100010001010000"
},
/* 28*/ { BARCODE_DAFT, "DAFTTFADFATDTATFT", 0, 3, 33, "Verified manually against tec-it",
"001010000010100010100000001000100"
"101010101010101010101010101010101"
"100010000010001010000010000000100"
},
};
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 data_size = ARRAY_SIZE(data);
int i, length, ret;
struct zint_symbol *symbol = NULL;
char escaped[1024];
char bwipp_buf[8192];
char bwipp_msg[1024];
int do_bwipp = (debug & ZINT_DEBUG_TEST_BWIPP) && testUtilHaveGhostscript(); /* Only do BWIPP test if asked, too slow otherwise */
testStartSymbol("test_encode", &symbol);
for (i = 0; i < data_size; i++) {
if (testContinue(p_ctx, i)) continue;
symbol = ZBarcode_Create();
assert_nonnull(symbol, "Symbol not created\n");
length = testUtilSetSymbol(symbol, data[i].symbology, -1 /*input_mode*/, -1 /*eci*/, -1 /*option_1*/, -1, -1, -1 /*output_options*/, data[i].data, -1, debug);
ret = ZBarcode_Encode(symbol, (unsigned char *) data[i].data, length);
assert_equal(ret, data[i].ret, "i:%d ZBarcode_Encode ret %d != %d (%s)\n", i, ret, data[i].ret, symbol->errtxt);
if (p_ctx->generate) {
printf(" /*%3d*/ { %s, \"%s\", %s, %d, %d, \"%s\",\n",
i, testUtilBarcodeName(data[i].symbology), testUtilEscape(data[i].data, length, escaped, sizeof(escaped)),
testUtilErrorName(data[i].ret), symbol->rows, symbol->width, data[i].comment);
testUtilModulesPrint(symbol, " ", "\n");
printf(" },\n");
} else {
if (ret < ZINT_ERROR) {
int width, row;
assert_equal(symbol->rows, data[i].expected_rows, "i:%d symbol->rows %d != %d (%s)\n", i, symbol->rows, data[i].expected_rows, data[i].data);
assert_equal(symbol->width, data[i].expected_width, "i:%d symbol->width %d != %d (%s)\n", i, symbol->width, data[i].expected_width, data[i].data);
ret = testUtilModulesCmp(symbol, data[i].expected, &width, &row);
assert_zero(ret, "i:%d testUtilModulesCmp ret %d != 0 width %d row %d (%s)\n", i, ret, width, row, data[i].data);
if (do_bwipp && testUtilCanBwipp(i, symbol, -1, -1, -1, debug)) {
ret = testUtilBwipp(i, symbol, -1, -1, -1, data[i].data, length, NULL, bwipp_buf, sizeof(bwipp_buf), NULL);
assert_zero(ret, "i:%d %s testUtilBwipp ret %d != 0\n", i, testUtilBarcodeName(symbol->symbology), ret);
ret = testUtilBwippCmp(symbol, bwipp_msg, bwipp_buf, data[i].expected);
assert_zero(ret, "i:%d %s testUtilBwippCmp %d != 0 %s\n actual: %s\nexpected: %s\n",
i, testUtilBarcodeName(symbol->symbology), ret, bwipp_msg, bwipp_buf, data[i].expected);
}
}
}
ZBarcode_Delete(symbol);
}
testFinish();
}
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
#include <time.h>
#define TEST_PERF_ITER_MILLES 10
#define TEST_PERF_ITERATIONS (TEST_PERF_ITER_MILLES * 1000)
#define TEST_PERF_TIME(arg) ((arg) * 1000.0 / CLOCKS_PER_SEC)
/* Not a real test, just performance indicator */
static void test_perf(const testCtx *const p_ctx) {
int debug = p_ctx->debug;
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
struct item {
int symbology;
int option_2;
char *data;
int ret;
int expected_rows;
int expected_width;
char *comment;
};
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 struct item data[] = {
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
/* 0*/ { BARCODE_POSTNET, -1, "12345678901", 0, 2, 123, "POSTNET 11" },
/* 1*/ { BARCODE_PLANET, -1, "1234567890123", 0, 2, 143, "PLANET 13" },
/* 2*/ { BARCODE_KOREAPOST, -1, "123456", 0, 1, 167, "KOREAPOST 6" },
};
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 data_size = ARRAY_SIZE(data);
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, length, ret;
struct zint_symbol *symbol;
clock_t start;
clock_t total_create = 0, total_encode = 0, total_buffer = 0, total_buf_inter = 0, total_print = 0;
clock_t diff_create, diff_encode, diff_buffer, diff_buf_inter, diff_print;
int comment_max = 0;
if (!(debug & ZINT_DEBUG_TEST_PERFORMANCE)) { /* -d 256 */
return;
}
for (i = 0; i < data_size; i++) if ((int) strlen(data[i].comment) > comment_max) comment_max = (int) strlen(data[i].comment);
printf("Iterations %d\n", TEST_PERF_ITERATIONS);
for (i = 0; i < data_size; i++) {
int j;
if (testContinue(p_ctx, i)) continue;
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
diff_create = diff_encode = diff_buffer = diff_buf_inter = diff_print = 0;
for (j = 0; j < TEST_PERF_ITERATIONS; j++) {
start = clock();
symbol = ZBarcode_Create();
diff_create += clock() - start;
assert_nonnull(symbol, "Symbol not created\n");
length = testUtilSetSymbol(symbol, data[i].symbology, DATA_MODE, -1 /*eci*/, -1 /*option_1*/, data[i].option_2, -1, -1 /*output_options*/, data[i].data, -1, debug);
start = clock();
ret = ZBarcode_Encode(symbol, (unsigned char *) data[i].data, length);
diff_encode += clock() - start;
assert_equal(ret, data[i].ret, "i:%d ZBarcode_Encode ret %d != %d (%s)\n", i, ret, data[i].ret, symbol->errtxt);
assert_equal(symbol->rows, data[i].expected_rows, "i:%d symbol->rows %d != %d (%s)\n", i, symbol->rows, data[i].expected_rows, data[i].data);
assert_equal(symbol->width, data[i].expected_width, "i:%d symbol->width %d != %d (%s)\n", i, symbol->width, data[i].expected_width, data[i].data);
start = clock();
ret = ZBarcode_Buffer(symbol, 0 /*rotate_angle*/);
diff_buffer += clock() - start;
assert_zero(ret, "i:%d ZBarcode_Buffer ret %d != 0 (%s)\n", i, ret, symbol->errtxt);
symbol->output_options |= OUT_BUFFER_INTERMEDIATE;
start = clock();
ret = ZBarcode_Buffer(symbol, 0 /*rotate_angle*/);
diff_buf_inter += clock() - start;
assert_zero(ret, "i:%d ZBarcode_Buffer OUT_BUFFER_INTERMEDIATE ret %d != 0 (%s)\n", i, ret, symbol->errtxt);
symbol->output_options &= ~OUT_BUFFER_INTERMEDIATE; /* Undo */
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
start = clock();
ret = ZBarcode_Print(symbol, 0 /*rotate_angle*/);
diff_print += clock() - start;
assert_zero(ret, "i:%d ZBarcode_Print ret %d != 0 (%s)\n", i, ret, symbol->errtxt);
assert_zero(testUtilRemove(symbol->outfile), "i:%d testUtilRemove(%s) != 0\n", i, symbol->outfile);
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
ZBarcode_Delete(symbol);
}
printf("%*s: encode % 8gms, buffer % 8gms, buf_inter % 8gms, print % 8gms, create % 8gms\n", comment_max, data[i].comment,
TEST_PERF_TIME(diff_encode), TEST_PERF_TIME(diff_buffer), TEST_PERF_TIME(diff_buf_inter), TEST_PERF_TIME(diff_print), TEST_PERF_TIME(diff_create));
total_create += diff_create;
total_encode += diff_encode;
total_buffer += diff_buffer;
total_buf_inter += diff_buf_inter;
total_print += diff_print;
}
if (p_ctx->index == -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
printf("%*s: encode % 8gms, buffer % 8gms, buf_inter % 8gms, print % 8gms, create % 8gms\n", comment_max, "totals",
TEST_PERF_TIME(total_encode), TEST_PERF_TIME(total_buffer), TEST_PERF_TIME(total_buf_inter), TEST_PERF_TIME(total_print), TEST_PERF_TIME(total_create));
}
}
int main(int argc, char *argv[]) {
testFunction funcs[] = { /* name, func */
{ "test_large", test_large },
{ "test_koreapost", test_koreapost },
{ "test_japanpost", test_japanpost },
{ "test_input", test_input },
{ "test_encode", test_encode },
{ "test_perf", test_perf },
};
testRun(argc, argv, funcs, ARRAY_SIZE(funcs));
testReport();
return 0;
}
/* vim: set ts=4 sw=4 et : */