zint/backend/png.c

334 lines
12 KiB
C
Raw Normal View History

2016-02-17 23:52:48 +13:00
/* png.c - Handles output to PNG file */
/*
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) 2009-2024 Robin Stuart <rstuart114@gmail.com>
2016-02-17 23:52:48 +13:00
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
2017-10-24 08:37:52 +13:00
1. Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
2016-02-17 23:52:48 +13:00
2. Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
2017-10-24 08:37:52 +13:00
documentation and/or other materials provided with the distribution.
2016-02-17 23:52:48 +13:00
3. Neither the name of the project nor the names of its contributors
may be used to endorse or promote products derived from this software
2017-10-24 08:37:52 +13:00
without specific prior written permission.
2016-02-17 23:52:48 +13:00
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
2017-10-24 08:37:52 +13:00
OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
2016-02-17 23:52:48 +13:00
SUCH DAMAGE.
*/
/* SPDX-License-Identifier: BSD-3-Clause */
2016-02-17 23:52:48 +13:00
#ifndef ZINT_NO_PNG
#include <errno.h>
- API: add new zint_symbol `dpmm` field for output resolution (BMP/ EMF/PCX/PNG/TIF only, i.e. excluding EPS, GIF & SVG) - Add support for specifying scale by X-dimension and resolution with new option `--scalexdimdp` for CLI/Tcl & new API function `ZBarcode_Scale_From_XdimDp()` (+ `ZBarcode_XdimDp_From_Scale()` & `ZBarcode_Default_Xdim()`) and new GUI popup; manual: document - BMP/EMF/PCX/PNG/TIF: use new `dpmm` resolution field (for EMF following Inkscape) - backend_qt: add `dpmm()`, `vectorWidth()`, `vectorHeight()`, `noPng()`, `getVersion()`, `takesGS1AIData()`, & `XdimDp` stuff incl. new `QZintXdimDp` struct for passing around scale vars & use in `getAsCLI()`; add comments - Raise `scale` limit to 200 (from 100) to allow for large dpmm - output: create directories & subdirectories as necessary for output path using new function `out_fopen()` and use in BMP/EMF/ EPS/GIF/PCX/PNG/SVG/TIF - DPLEIT/DPIDENT: format HRT according to (incomplete) documentation, and set default height to 72X (from 50X) - CODE128B renamed to CODE128AB as can use subsets A and/or B - CODABAR: fix minimum height calc - EMF: fix indexing of handles (zero-based not 1-based) - GUI: fix symbology zap (previous technique of clearing and re-loading settings without doing a sync no longer works); fix UPCEAN guard descent enable - MAILMARK: better error message if input < 14 characters - GUI: add "Default" button for DAFT tracker ratio & enable/disable various default buttons; use new `takesGS1AIData()` to enable/disable GS1-specific checkboxes - CLI: use new `validate_float()` to parse float options (7 significant digits allowed only, no scientific notation) - DATAMATRIX/GRIDMATRIX/PDF417/QR/ULTRA: micro-optimize structapp ID parse - library/CLI: fiddle with static asserts (make CHAR_BIT sensitive, supposedly) - win32/README: update building libpng (assembly removed) - README.linux: document incompatibility of Qt6 >= 6.3 - manual: expand Barcode Studio waffle - test suite: change range separator to hyphen and allow multiple excludes
2022-12-03 10:39:01 +13:00
#include <math.h>
2016-02-17 23:52:48 +13:00
#include <stdio.h>
#include <png.h>
#include <zlib.h>
#include <setjmp.h>
#include "common.h"
#include "filemem.h"
- API: add new zint_symbol `dpmm` field for output resolution (BMP/ EMF/PCX/PNG/TIF only, i.e. excluding EPS, GIF & SVG) - Add support for specifying scale by X-dimension and resolution with new option `--scalexdimdp` for CLI/Tcl & new API function `ZBarcode_Scale_From_XdimDp()` (+ `ZBarcode_XdimDp_From_Scale()` & `ZBarcode_Default_Xdim()`) and new GUI popup; manual: document - BMP/EMF/PCX/PNG/TIF: use new `dpmm` resolution field (for EMF following Inkscape) - backend_qt: add `dpmm()`, `vectorWidth()`, `vectorHeight()`, `noPng()`, `getVersion()`, `takesGS1AIData()`, & `XdimDp` stuff incl. new `QZintXdimDp` struct for passing around scale vars & use in `getAsCLI()`; add comments - Raise `scale` limit to 200 (from 100) to allow for large dpmm - output: create directories & subdirectories as necessary for output path using new function `out_fopen()` and use in BMP/EMF/ EPS/GIF/PCX/PNG/SVG/TIF - DPLEIT/DPIDENT: format HRT according to (incomplete) documentation, and set default height to 72X (from 50X) - CODE128B renamed to CODE128AB as can use subsets A and/or B - CODABAR: fix minimum height calc - EMF: fix indexing of handles (zero-based not 1-based) - GUI: fix symbology zap (previous technique of clearing and re-loading settings without doing a sync no longer works); fix UPCEAN guard descent enable - MAILMARK: better error message if input < 14 characters - GUI: add "Default" button for DAFT tracker ratio & enable/disable various default buttons; use new `takesGS1AIData()` to enable/disable GS1-specific checkboxes - CLI: use new `validate_float()` to parse float options (7 significant digits allowed only, no scientific notation) - DATAMATRIX/GRIDMATRIX/PDF417/QR/ULTRA: micro-optimize structapp ID parse - library/CLI: fiddle with static asserts (make CHAR_BIT sensitive, supposedly) - win32/README: update building libpng (assembly removed) - README.linux: document incompatibility of Qt6 >= 6.3 - manual: expand Barcode Studio waffle - test suite: change range separator to hyphen and allow multiple excludes
2022-12-03 10:39:01 +13:00
#include "output.h"
/* Note using "wpng_" prefix not "png_" (except for `png_pixel_plot()`) to avoid clashing with libpng */
/* Note if change this need to change "backend/tests/test_png.c" definition also */
struct wpng_error_type {
struct zint_symbol *symbol;
2016-02-17 23:52:48 +13:00
jmp_buf jmpbuf;
};
static void wpng_error_handler(png_structp png_ptr, png_const_charp msg) {
struct wpng_error_type *wpng_error_ptr;
2016-02-17 23:52:48 +13:00
wpng_error_ptr = (struct wpng_error_type *) png_get_error_ptr(png_ptr);
if (wpng_error_ptr == NULL) {
/* we are completely hosed now */
fprintf(stderr, "Error 636: libpng error: %s\n", msg ? msg : "<NULL>");
fprintf(stderr, "Error 637: jmpbuf not recoverable, terminating\n");
2016-02-17 23:52:48 +13:00
fflush(stderr);
return; /* libpng will call abort() */
2016-02-17 23:52:48 +13:00
}
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
errtxtf(0, wpng_error_ptr->symbol, 635, "libpng error: %s", msg ? msg : "<NULL>");
longjmp(wpng_error_ptr->jmpbuf, 1);
2016-02-17 23:52:48 +13:00
}
#ifdef ZINT_TEST /* Wrapper for direct testing */
INTERNAL void wpng_error_handler_test(png_structp png_ptr, png_const_charp msg) {
EAN/UPC: add quiet zone indicators option (API `output_options` `EANUPC_GUARD_WHITESPACE`, CLI `--guardwhitespace`) (ticket #287) EAN-2/EAN-5: HRT now at top instead of at bottom for standalones, following BWIPP CLI: batch mode: don't close input if stdin EAN/UPC: fix excess 1X to right of add-ons Composites: fix excess whitespace; fix quiet zone calcs to allow for linear shifting CLI: use own (Wine) version of `CommandLineToArgvW()` to avoid loading "shell32.dll" Move "font.h" -> "raster_font.h" EPS/SVG: use new `out_putsf()` func to output floats, avoiding trailing zeroes & locale dependency EPS: simplify "TR" formula SVG: change font from "Helvetica, sans serif" to "OCR-B, monospace"; use single "<path>" instead of multiple "<rect>"s to draw boxes (reduces file size) Add `EMBED_VECTOR_FONT` to `output_options` (CLI `--embedfont`) to enable embedding of font in vector output - currently only for SVG output of EAN/UPC GUI: use "OCR-B" font for EAN/UPC and "Arimo" for all others (was "Helvetica" for both); paint background of screen preview light grey so as contrasts with whitespace and quiet zones EMF: prefix funcs with `emf_`; simplify string `halign` handling large: rename `large_int` -> `large_uint` CODE128/common: move `c128_hrt_cpy_iso8859_1()` to `hrt_cpy_iso8859_1()` and add `ZINT_WARN_HRT_TRUNCATED` warning (for future use) Various symbologies: replace `printf()` with `fputs()` (symbol->debug) QRCODE: better assert(), removing a NOLINT (2 left) CLI: add some more barcode synonyms for DBAR common: various fiddlings CMake: don't include png.c unless ZINT_USE_PNG (avoids clang warning)
2023-06-12 12:25:55 +12:00
wpng_error_handler(png_ptr, msg);
}
#endif
/* libpng write callback */
static void wpng_write(png_structp png_ptr, png_bytep ptr, size_t size) {
struct filemem *fmp = (struct filemem *) png_get_io_ptr(png_ptr);
(void) fm_write(ptr, 1, size, fmp);
}
/* libpng flush callback */
static void wpng_flush(png_structp png_ptr) {
struct filemem *fmp = (struct filemem *) png_get_io_ptr(png_ptr);
(void) fm_flush(fmp);
}
/* Guestimate best compression strategy */
static int wpng_guess_compression_strategy(struct zint_symbol *symbol, const unsigned char *pixelbuf) {
(void)pixelbuf;
/* TODO: Do properly */
/* It seems the best choice for typical barcode pngs is one of Z_DEFAULT_STRATEGY and Z_FILTERED */
/* Some guesses */
if (symbol->symbology == BARCODE_MAXICODE) {
return Z_DEFAULT_STRATEGY;
}
if (symbol->symbology == BARCODE_AZTEC && symbol->bitmap_width <= 30) {
return Z_DEFAULT_STRATEGY;
}
/* Z_FILTERED seems to work better for slightly more barcodes+data so default to that */
return Z_FILTERED;
}
INTERNAL int png_pixel_plot(struct zint_symbol *symbol, const unsigned char *pixelbuf) {
struct wpng_error_type wpng_error;
struct filemem fm;
struct filemem *const fmp = &fm;
png_structp png_ptr;
png_infop info_ptr;
int i;
int row, column;
png_color bg, fg;
unsigned char bg_alpha, fg_alpha;
unsigned char map[128];
png_color palette[32];
int num_palette;
unsigned char trans_alpha[32];
int num_trans; /* Note initialize below to avoid gcc -Wclobbered warning due to `longjmp()` */
int bit_depth;
int compression_strategy;
const unsigned char *pb;
unsigned char *outdata = (unsigned char *) z_alloca(symbol->bitmap_width);
2016-02-17 23:52:48 +13:00
wpng_error.symbol = symbol;
(void) out_colour_get_rgb(symbol->fgcolour, &fg.red, &fg.green, &fg.blue, &fg_alpha);
(void) out_colour_get_rgb(symbol->bgcolour, &bg.red, &bg.green, &bg.blue, &bg_alpha);
num_trans = 0;
if (symbol->symbology == BARCODE_ULTRA) {
static const unsigned char ultra_chars[8] = { 'W', 'C', 'B', 'M', 'R', 'Y', 'G', 'K' };
for (i = 0; i < 8; i++) {
map[ultra_chars[i]] = i;
out_colour_char_to_rgb(ultra_chars[i], &palette[i].red, &palette[i].green, &palette[i].blue);
if (fg_alpha != 0xff) {
trans_alpha[i] = fg_alpha;
}
}
num_palette = 8;
if (fg_alpha != 0xff) {
num_trans = 8;
}
/* For Ultracode, have foreground only if have bind/box */
if (symbol->border_width > 0 && (symbol->output_options & (BARCODE_BIND | BARCODE_BOX | BARCODE_BIND_TOP))) {
/* Check whether can re-use black */
if (fg.red == 0 && fg.green == 0 && fg.blue == 0) {
map['1'] = 7; /* Re-use black */
} else {
map['1'] = num_palette;
palette[num_palette++] = fg;
if (fg_alpha != 0xff) {
trans_alpha[num_trans++] = fg_alpha;
}
}
}
/* For Ultracode, have background only if have whitespace/quiet zones */
if (symbol->whitespace_width > 0 || symbol->whitespace_height > 0
|| ((symbol->output_options & BARCODE_QUIET_ZONES)
&& !(symbol->output_options & BARCODE_NO_QUIET_ZONES))) {
/* Check whether can re-use white */
if (bg.red == 0xff && bg.green == 0xff && bg.blue == 0xff && bg_alpha == fg_alpha) {
map['0'] = 0; /* Re-use white */
} else {
if (bg_alpha == 0xff || fg_alpha != 0xff) {
/* No alpha or have foreground alpha - add to end */
map['0'] = num_palette;
palette[num_palette++] = bg;
} else {
/* Alpha and no foreground alpha - add to front & move white to end */
png_color white = palette[0]; /* Take copy */
map['0'] = 0;
palette[0] = bg;
map['W'] = num_palette;
palette[num_palette++] = white;
}
if (bg_alpha != 0xff) {
trans_alpha[num_trans++] = bg_alpha;
}
}
}
} else {
int bg_idx = 0, fg_idx = 1;
/* Do alphas first so can swop indexes if background not alpha */
if (bg_alpha != 0xff) {
trans_alpha[num_trans++] = bg_alpha;
}
if (fg_alpha != 0xff) {
trans_alpha[num_trans++] = fg_alpha;
if (num_trans == 1) {
/* Only foreground has alpha so swop indexes - saves a byte! */
bg_idx = 1;
fg_idx = 0;
}
}
map['0'] = bg_idx;
palette[bg_idx] = bg;
map['1'] = fg_idx;
palette[fg_idx] = fg;
num_palette = 2;
}
if (num_palette <= 2) {
bit_depth = 1;
} else {
bit_depth = 4;
}
/* Open output file in binary mode */
if (!fm_open(fmp, symbol, "wb")) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxtf(ZINT_ERROR_FILE_ACCESS, symbol, 632, "Could not open PNG output file (%1$d: %2$s)", fmp->err,
strerror(fmp->err));
}
/* Set up error handling routine as proc() above */
png_ptr = png_create_write_struct(PNG_LIBPNG_VER_STRING, &wpng_error, wpng_error_handler, NULL);
if (!png_ptr) {
(void) fm_close(fmp, symbol);
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxt(ZINT_ERROR_MEMORY, symbol, 633, "Insufficient memory for PNG write structure buffer");
}
info_ptr = png_create_info_struct(png_ptr);
if (!info_ptr) {
png_destroy_write_struct(&png_ptr, NULL);
(void) fm_close(fmp, symbol);
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxt(ZINT_ERROR_MEMORY, symbol, 634, "Insufficient memory for PNG info structure buffer");
}
/* catch jumping here */
if (setjmp(wpng_error.jmpbuf)) {
png_destroy_write_struct(&png_ptr, &info_ptr);
(void) fm_close(fmp, symbol);
return ZINT_ERROR_MEMORY;
}
/* Set our output functions */
png_set_write_fn(png_ptr, fmp, wpng_write, wpng_flush);
/* set compression */
png_set_compression_level(png_ptr, 9);
/* Compression strategy can make a difference */
compression_strategy = wpng_guess_compression_strategy(symbol, pixelbuf);
if (compression_strategy != Z_DEFAULT_STRATEGY) {
png_set_compression_strategy(png_ptr, compression_strategy);
}
- API: add new zint_symbol `dpmm` field for output resolution (BMP/ EMF/PCX/PNG/TIF only, i.e. excluding EPS, GIF & SVG) - Add support for specifying scale by X-dimension and resolution with new option `--scalexdimdp` for CLI/Tcl & new API function `ZBarcode_Scale_From_XdimDp()` (+ `ZBarcode_XdimDp_From_Scale()` & `ZBarcode_Default_Xdim()`) and new GUI popup; manual: document - BMP/EMF/PCX/PNG/TIF: use new `dpmm` resolution field (for EMF following Inkscape) - backend_qt: add `dpmm()`, `vectorWidth()`, `vectorHeight()`, `noPng()`, `getVersion()`, `takesGS1AIData()`, & `XdimDp` stuff incl. new `QZintXdimDp` struct for passing around scale vars & use in `getAsCLI()`; add comments - Raise `scale` limit to 200 (from 100) to allow for large dpmm - output: create directories & subdirectories as necessary for output path using new function `out_fopen()` and use in BMP/EMF/ EPS/GIF/PCX/PNG/SVG/TIF - DPLEIT/DPIDENT: format HRT according to (incomplete) documentation, and set default height to 72X (from 50X) - CODE128B renamed to CODE128AB as can use subsets A and/or B - CODABAR: fix minimum height calc - EMF: fix indexing of handles (zero-based not 1-based) - GUI: fix symbology zap (previous technique of clearing and re-loading settings without doing a sync no longer works); fix UPCEAN guard descent enable - MAILMARK: better error message if input < 14 characters - GUI: add "Default" button for DAFT tracker ratio & enable/disable various default buttons; use new `takesGS1AIData()` to enable/disable GS1-specific checkboxes - CLI: use new `validate_float()` to parse float options (7 significant digits allowed only, no scientific notation) - DATAMATRIX/GRIDMATRIX/PDF417/QR/ULTRA: micro-optimize structapp ID parse - library/CLI: fiddle with static asserts (make CHAR_BIT sensitive, supposedly) - win32/README: update building libpng (assembly removed) - README.linux: document incompatibility of Qt6 >= 6.3 - manual: expand Barcode Studio waffle - test suite: change range separator to hyphen and allow multiple excludes
2022-12-03 10:39:01 +13:00
if (symbol->dpmm) {
int resolution = (int) roundf(stripf(symbol->dpmm * 1000.0f)); /* pixels per metre */
png_set_pHYs(png_ptr, info_ptr, resolution, resolution, PNG_RESOLUTION_METER);
}
/* set Header block */
png_set_IHDR(png_ptr, info_ptr, symbol->bitmap_width, symbol->bitmap_height,
bit_depth, PNG_COLOR_TYPE_PALETTE, PNG_INTERLACE_NONE,
PNG_COMPRESSION_TYPE_DEFAULT, PNG_FILTER_TYPE_DEFAULT);
png_set_PLTE(png_ptr, info_ptr, palette, num_palette);
if (num_trans) {
png_set_tRNS(png_ptr, info_ptr, trans_alpha, num_trans, NULL);
}
/* write all chunks up to (but not including) first IDAT */
png_write_info(png_ptr, info_ptr);
/* Pixel Plotting */
pb = pixelbuf;
if (bit_depth == 1) {
for (row = 0; row < symbol->bitmap_height; row++) {
Performance improvements for linear encoding and raster output - use fixed-length string tables (mostly) instead of (char *) pointer ones (saves ~40K) - re-use C128Table for CODABLOCKF and CODE16K (required removal of Stop character and extra CODE16K-only entry) - use pointer to destination and copy (memcpy/strcpy(), bin_append_posn()) instead of concatenating (strcat()) (mostly) - replace last remaining bin_append()s with bin_append_posn(); bin_append() removed - add length arg to toupper() and expand() (avoids strlen()) - change is_sane() to use table-based flags (avoids an iteration) - rename lookup() to is_sane_lookup() and change to check and return posns and use in pointer to destination loops (avoids strcat()s) - remove special case PHARMA in expand() (dealt with in pharma()) - make #define SILVER/CALCIUM/TECHNETIUM/KRSET etc static strings - replace strchr() -> posn() - CODE128: populate destination once in checksum loop; re-use and export some more routines (c128_set_a/b/c(), c128_put_in_set()) for sharing; prefix defines (SHIFTA -> C128_SHIFTA etc) and existing exported routines - use factor XOR toggle trick in checksum calcs (avoids branch) - raster.c: fill out single 1-pixel row and copy using new draw_bar_line(), copy_bar_line() routines; similarly in buffer_plot compare previous line & copy if same (same technique as used to improve non-half-integer scaling, significant performance increase, (c) codemonkey82); also done for PNG (BMP/GIF/PCX/TIFF not done) - raster/vector/output.c: shorten "output_" prefix -> "out_"; sync vector to other raster changes to try to keep source files similar - 2of5.c: prefix "c25_" JAPANPOST: return error if input data truncated (backward incompatible) DAFT: max chars 50 -> 100 common.c: istwodigit() -> is_twodigit() common.c/emf.c/output.c: use some further stripf()s (MSVC6 float variations) library.c: new check_output_args() helper zint.h: add BARCODE_LAST marker and use in library.c QRCODE: remove a NOLINT (requires clang-tidy-13), one remaining CMake: separate no-optimize from ZINT_DEBUG into new ZINT_NOOPT option
2021-10-21 11:05:30 +13:00
if (row && memcmp(pb, pb - symbol->bitmap_width, symbol->bitmap_width) == 0) {
pb += symbol->bitmap_width;
} else {
unsigned char *image_data = outdata;
for (column = 0; column < symbol->bitmap_width; column += 8, image_data++) {
unsigned char byte = 0;
for (i = 0; i < 8 && column + i < symbol->bitmap_width; i++, pb++) {
byte |= map[*pb] << (7 - i);
}
*image_data = byte;
}
}
/* write row contents to file */
png_write_row(png_ptr, outdata);
}
} else { /* Bit depth 4 */
for (row = 0; row < symbol->bitmap_height; row++) {
Performance improvements for linear encoding and raster output - use fixed-length string tables (mostly) instead of (char *) pointer ones (saves ~40K) - re-use C128Table for CODABLOCKF and CODE16K (required removal of Stop character and extra CODE16K-only entry) - use pointer to destination and copy (memcpy/strcpy(), bin_append_posn()) instead of concatenating (strcat()) (mostly) - replace last remaining bin_append()s with bin_append_posn(); bin_append() removed - add length arg to toupper() and expand() (avoids strlen()) - change is_sane() to use table-based flags (avoids an iteration) - rename lookup() to is_sane_lookup() and change to check and return posns and use in pointer to destination loops (avoids strcat()s) - remove special case PHARMA in expand() (dealt with in pharma()) - make #define SILVER/CALCIUM/TECHNETIUM/KRSET etc static strings - replace strchr() -> posn() - CODE128: populate destination once in checksum loop; re-use and export some more routines (c128_set_a/b/c(), c128_put_in_set()) for sharing; prefix defines (SHIFTA -> C128_SHIFTA etc) and existing exported routines - use factor XOR toggle trick in checksum calcs (avoids branch) - raster.c: fill out single 1-pixel row and copy using new draw_bar_line(), copy_bar_line() routines; similarly in buffer_plot compare previous line & copy if same (same technique as used to improve non-half-integer scaling, significant performance increase, (c) codemonkey82); also done for PNG (BMP/GIF/PCX/TIFF not done) - raster/vector/output.c: shorten "output_" prefix -> "out_"; sync vector to other raster changes to try to keep source files similar - 2of5.c: prefix "c25_" JAPANPOST: return error if input data truncated (backward incompatible) DAFT: max chars 50 -> 100 common.c: istwodigit() -> is_twodigit() common.c/emf.c/output.c: use some further stripf()s (MSVC6 float variations) library.c: new check_output_args() helper zint.h: add BARCODE_LAST marker and use in library.c QRCODE: remove a NOLINT (requires clang-tidy-13), one remaining CMake: separate no-optimize from ZINT_DEBUG into new ZINT_NOOPT option
2021-10-21 11:05:30 +13:00
if (row && memcmp(pb, pb - symbol->bitmap_width, symbol->bitmap_width) == 0) {
pb += symbol->bitmap_width;
} else {
unsigned char *image_data = outdata;
for (column = 0; column < symbol->bitmap_width; column += 2, image_data++) {
unsigned char byte = map[*pb++] << 4;
if (column + 1 < symbol->bitmap_width) {
byte |= map[*pb++];
}
*image_data = byte;
}
}
/* write row contents to file */
png_write_row(png_ptr, outdata);
}
}
/* End the file */
png_write_end(png_ptr, NULL);
/* make sure we have disengaged */
png_destroy_write_struct(&png_ptr, &info_ptr);
if (fm_error(fmp)) {
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
errtxtf(0, symbol, 638, "Incomplete write of PNG output (%1$d: %2$s)", fmp->err, strerror(fmp->err));
(void) fm_close(fmp, symbol);
return ZINT_ERROR_FILE_WRITE;
}
if (!fm_close(fmp, symbol)) {
CODE128: reduce extended latch cut-off from 5 to 4 for better encodation in certain cases (and no pessimizations found so far), props lyngklip (BWIPP); fix extended char latching when exactly 3 extended chars at end; count code set C (not digits) in loop deciding when to shift/latch to extended for better estimate AZTEC: return warning if ECC < 5% (due to bit-stuffing when version given); return error if > 22 layers (Zint 26) for Reader Initialisation symbol requested for better error message AZTEC/HANXIN/QRCODE: consolidate different ECC data size tables into one indexed by ECC DBAR_EXP: check for reduced length <= 77 up front for better error message HANXIN: use `malloc()` rather than `z_alloca()` for large binary array QRCODE: `ecc_level` now 0-based (not 1-based) MICROQR: consolidate different version end routines into one `microqr_end()` and use new `microqr_data` table to simplify code MICROPDF417: use table for max codewords per column library: centralize all error messages using new `errtxt()`, `errtxtf()`, `errtxt_adj()` funcs that protect `symbol->errtxt` from overflow, & try to make error messages more consistent thru-out, adding more feedback info to many, & use positional args "%n$" in prep for l10n (maybe); `is_sane/is_sane_lookup()` -> `not_sane/not_sane_lookup()`, returning 1-based position (zero on failure) instead of bool; `long` ints -> plain `int` (except those dealing with `ftell()`, `fread()` etc) as depend on int being 32-bits already GUI: in "grpDATF.ui" use "PlainText" rather than "RichText" for tracker ratio examples as height of text messing up sometimes manual: clarify Codablock-F length maximum & add examples docs: README: pandoc 3.5, Ubuntu 24.04 CMake: use "-Wpedantic" for Clang only as GNU complains about `errtxtf()` positional args "%n$"
2024-10-28 10:33:33 +13:00
return errtxtf(ZINT_ERROR_FILE_WRITE, symbol, 960, "Failure on closing PNG output file (%1$d: %2$s)",
fmp->err, strerror(fmp->err));
}
return 0;
2016-02-17 23:52:48 +13:00
}
/* vim: set ts=4 sw=4 et : */
#else
#if defined(__clang__)
/* Suppresses clang-tidy-18 "clang-diagnostic-empty-translation-unit" */
typedef int wpng_make_clang_tidy_compilers_happy;
#endif
#endif /* ZINT_NO_PNG */