Update bundled libpng to version 1.6.44

[ChangeLog][Third-Party Code] libpng was updated to version 1.6.44

Change-Id: Ib208635687645101cd2d7fb4e5bf069088c24f86
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
(cherry picked from commit 685a6c2)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit c36b24b)

src/3rdparty/libpng/ANNOUNCE

@@ -1,5 +1,5 @@
-libpng 1.6.43 - February 23, 2024
-=================================
+libpng 1.6.44 - September 12, 2024
+==================================
 
 This is a public release of libpng, intended for use in production code.
 
@@ -9,13 +9,13 @@ Files available for download
 
 Source files with LF line endings (for Unix/Linux):
 
- * libpng-1.6.43.tar.xz (LZMA-compressed, recommended)
- * libpng-1.6.43.tar.gz (deflate-compressed)
+ * libpng-1.6.44.tar.xz (LZMA-compressed, recommended)
+ * libpng-1.6.44.tar.gz (deflate-compressed)
 
 Source files with CRLF line endings (for Windows):
 
- * lpng1643.7z (LZMA-compressed, recommended)
- * lpng1643.zip (deflate-compressed)
+ * lpng1644.7z (LZMA-compressed, recommended)
+ * lpng1644.zip (deflate-compressed)
 
 Other information:
 
@@ -25,36 +25,29 @@ Other information:
  * TRADEMARK.md
 
 
-Changes from version 1.6.42 to version 1.6.43
+Changes from version 1.6.43 to version 1.6.44
 ---------------------------------------------
 
- * Fixed the row width check in png_check_IHDR().
-   This corrected a bug that was specific to the 16-bit platforms,
-   and removed a spurious compiler warning from the 64-bit builds.
-   (Reported by Jacek Caban; fixed by John Bowler)
- * Added eXIf chunk support to the push-mode reader in pngpread.c.
-   (Contributed by Chris Blume)
- * Added contrib/pngexif for the benefit of the users who would like
-   to inspect the content of eXIf chunks.
- * Added contrib/conftest/basic.dfa, a basic build-time configuration.
+ * Hardened calculations in chroma handling to prevent overflows, and
+   relaxed a constraint in cHRM validation to accomodate the standard
+   ACES AP1 set of color primaries.
    (Contributed by John Bowler)
- * Fixed a preprocessor condition in pngread.c that broke build-time
-   configurations like contrib/conftest/pngcp.dfa.
-   (Contributed by John Bowler)
- * Added CMake build support for LoongArch LSX.
-   (Contributed by GuXiWei)
- * Fixed a CMake build error that occurred under a peculiar state of the
-   dependency tree. This was a regression introduced in libpng-1.6.41.
-   (Contributed by Dan Rosser)
- * Marked the installed libpng headers as system headers in CMake.
-   (Contributed by Benjamin Buch)
- * Updated the build support for RISCOS.
-   (Contributed by Cameron Cawley)
- * Updated the makefiles to allow cross-platform builds to initialize
-   conventional make variables like AR and ARFLAGS.
- * Added various improvements to the CI scripts in areas like version
-   consistency verification and text linting.
- * Added version consistency verification to pngtest.c also.
+ * Removed the ASM implementation of ARM Neon optimizations and updated
+   the build accordingly. Only the remaining C implementation shall be
+   used from now on, thus ensuring the support of the PAC/BTI security
+   features on ARM64.
+   (Contributed by Ross Burton and John Bowler)
+ * Fixed the pickup of the PNG_HARDWARE_OPTIMIZATIONS option in the
+   CMake build on FreeBSD/amd64. This is an important performance fix
+   on this platform.
+ * Applied various fixes and improvements to the CMake build.
+   (Contributed by Eric Riff, Benjamin Buch and Erik Scholz)
+ * Added fuzzing targets for the simplified read API.
+   (Contributed by Mikhail Khachayants)
+ * Fixed a build error involving pngtest.c under a custom config.
+   This was a regression introduced in a code cleanup in libpng-1.6.43.
+   (Contributed by Ben Wagner)
+ * Fixed and improved the config files for AppVeyor CI and Travis CI.
 
 
 Send comments/corrections/commendations to png-mng-implement at lists.sf.net.

src/3rdparty/libpng/CHANGES

@@ -6196,6 +6196,28 @@ Version 1.6.43 [February 23, 2024]
     consistency verification and text linting.
   Added version consistency verification to pngtest.c also.
 
+Version 1.6.44 [September 12, 2024]
+  Hardened calculations in chroma handling to prevent overflows, and
+    relaxed a constraint in cHRM validation to accomodate the standard
+    ACES AP1 set of color primaries.
+    (Contributed by John Bowler)
+  Removed the ASM implementation of ARM Neon optimizations and updated
+    the build accordingly. Only the remaining C implementation shall be
+    used from now on, thus ensuring the support of the PAC/BTI security
+    features on ARM64.
+    (Contributed by Ross Burton and John Bowler)
+  Fixed the pickup of the PNG_HARDWARE_OPTIMIZATIONS option in the
+    CMake build on FreeBSD/amd64. This is an important performance fix
+    on this platform.
+  Applied various fixes and improvements to the CMake build.
+    (Contributed by Eric Riff, Benjamin Buch and Erik Scholz)
+  Added fuzzing targets for the simplified read API.
+    (Contributed by Mikhail Khachayants)
+  Fixed a build error involving pngtest.c under a custom config.
+    This was a regression introduced in a code cleanup in libpng-1.6.43.
+    (Contributed by Ben Wagner)
+  Fixed and improved the config files for AppVeyor CI and Travis CI.
+
 Send comments/corrections/commendations to png-mng-implement at lists.sf.net.
 Subscription is required; visit
 https://lists.sourceforge.net/lists/listinfo/png-mng-implement

src/3rdparty/libpng/README

@@ -1,4 +1,4 @@
-README for libpng version 1.6.43
+README for libpng version 1.6.44
 ================================
 
 See the note about version numbers near the top of `png.h`.

src/3rdparty/libpng/libpng-manual.txt

@@ -9,7 +9,7 @@ libpng-manual.txt - A description on how to use and modify libpng
 
  Based on:
 
- libpng version 1.6.36, December 2018, through 1.6.43 - February 2024
+ libpng version 1.6.36, December 2018, through 1.6.44 - September 2024
  Updated and distributed by Cosmin Truta
  Copyright (c) 2018-2024 Cosmin Truta
 

src/3rdparty/libpng/png.c

@@ -14,7 +14,7 @@
 #include "pngpriv.h"
 
 /* Generate a compiler error if there is an old png.h in the search path. */
-typedef png_libpng_version_1_6_43 Your_png_h_is_not_version_1_6_43;
+typedef png_libpng_version_1_6_44 Your_png_h_is_not_version_1_6_44;
 
 /* Tells libpng that we have already handled the first "num_bytes" bytes
  * of the PNG file signature.  If the PNG data is embedded into another
@@ -794,7 +794,7 @@ png_get_copyright(png_const_structrp png_ptr)
    return PNG_STRING_COPYRIGHT
 #else
    return PNG_STRING_NEWLINE \
-      "libpng version 1.6.43" PNG_STRING_NEWLINE \
+      "libpng version 1.6.44" PNG_STRING_NEWLINE \
       "Copyright (c) 2018-2024 Cosmin Truta" PNG_STRING_NEWLINE \
       "Copyright (c) 1998-2002,2004,2006-2018 Glenn Randers-Pehrson" \
       PNG_STRING_NEWLINE \
@@ -1203,6 +1203,24 @@ png_colorspace_sync(png_const_structrp png_ptr, png_inforp info_ptr)
 #endif /* GAMMA */
 
 #ifdef PNG_COLORSPACE_SUPPORTED
+static int
+png_safe_add(png_int_32 *addend0_and_result, png_int_32 addend1,
+      png_int_32 addend2) {
+   /* Safely add three integers.  Returns 0 on success, 1 on overlow.
+    * IMPLEMENTATION NOTE: ANSI requires signed overflow not to occur, therefore
+    * relying on addition of two positive values producing a negative one is not
+    * safe.
+    */
+   int addend0 = *addend0_and_result;
+   if (0x7fffffff - addend0 < addend1)
+      return 1;
+   addend0 += addend1;
+   if (0x7fffffff - addend1 < addend2)
+      return 1;
+   *addend0_and_result = addend0 + addend2;
+   return 0;
+}
+
 /* Added at libpng-1.5.5 to support read and write of true CIEXYZ values for
  * cHRM, as opposed to using chromaticities.  These internal APIs return
  * non-zero on a parameter error.  The X, Y and Z values are required to be
@@ -1211,38 +1229,52 @@ png_colorspace_sync(png_const_structrp png_ptr, png_inforp info_ptr)
 static int
 png_xy_from_XYZ(png_xy *xy, const png_XYZ *XYZ)
 {
-   png_int_32 d, dwhite, whiteX, whiteY;
+   png_int_32 d, dred, dgreen, dwhite, whiteX, whiteY;
 
-   d = XYZ->red_X + XYZ->red_Y + XYZ->red_Z;
+   /* 'd' in each of the blocks below is just X+Y+Z for each component,
+    * x, y and z are X,Y,Z/(X+Y+Z).
+    */
+   d = XYZ->red_X;
+   if (png_safe_add(&d, XYZ->red_Y, XYZ->red_Z))
+      return 1;
    if (png_muldiv(&xy->redx, XYZ->red_X, PNG_FP_1, d) == 0)
       return 1;
    if (png_muldiv(&xy->redy, XYZ->red_Y, PNG_FP_1, d) == 0)
       return 1;
-   dwhite = d;
+   dred = d;
    whiteX = XYZ->red_X;
    whiteY = XYZ->red_Y;
 
-   d = XYZ->green_X + XYZ->green_Y + XYZ->green_Z;
+   d = XYZ->green_X;
+   if (png_safe_add(&d, XYZ->green_Y, XYZ->green_Z))
+      return 1;
    if (png_muldiv(&xy->greenx, XYZ->green_X, PNG_FP_1, d) == 0)
       return 1;
    if (png_muldiv(&xy->greeny, XYZ->green_Y, PNG_FP_1, d) == 0)
       return 1;
-   dwhite += d;
+   dgreen = d;
    whiteX += XYZ->green_X;
    whiteY += XYZ->green_Y;
 
-   d = XYZ->blue_X + XYZ->blue_Y + XYZ->blue_Z;
+   d = XYZ->blue_X;
+   if (png_safe_add(&d, XYZ->blue_Y, XYZ->blue_Z))
+      return 1;
    if (png_muldiv(&xy->bluex, XYZ->blue_X, PNG_FP_1, d) == 0)
       return 1;
    if (png_muldiv(&xy->bluey, XYZ->blue_Y, PNG_FP_1, d) == 0)
       return 1;
-   dwhite += d;
    whiteX += XYZ->blue_X;
    whiteY += XYZ->blue_Y;
 
-   /* The reference white is simply the sum of the end-point (X,Y,Z) vectors,
-    * thus:
+   /* The reference white is simply the sum of the end-point (X,Y,Z) vectors so
+    * the fillowing calculates (X+Y+Z) of the reference white (media white,
+    * encoding white) itself:
     */
+   if (png_safe_add(&d, dred, dgreen))
+      return 1;
+
+   dwhite = d;
+
    if (png_muldiv(&xy->whitex, whiteX, PNG_FP_1, dwhite) == 0)
       return 1;
    if (png_muldiv(&xy->whitey, whiteY, PNG_FP_1, dwhite) == 0)
@@ -1257,20 +1289,6 @@ png_XYZ_from_xy(png_XYZ *XYZ, const png_xy *xy)
    png_fixed_point red_inverse, green_inverse, blue_scale;
    png_fixed_point left, right, denominator;
 
-   /* Check xy and, implicitly, z.  Note that wide gamut color spaces typically
-    * have end points with 0 tristimulus values (these are impossible end
-    * points, but they are used to cover the possible colors).  We check
-    * xy->whitey against 5, not 0, to avoid a possible integer overflow.
-    */
-   if (xy->redx   < 0 || xy->redx > PNG_FP_1) return 1;
-   if (xy->redy   < 0 || xy->redy > PNG_FP_1-xy->redx) return 1;
-   if (xy->greenx < 0 || xy->greenx > PNG_FP_1) return 1;
-   if (xy->greeny < 0 || xy->greeny > PNG_FP_1-xy->greenx) return 1;
-   if (xy->bluex  < 0 || xy->bluex > PNG_FP_1) return 1;
-   if (xy->bluey  < 0 || xy->bluey > PNG_FP_1-xy->bluex) return 1;
-   if (xy->whitex < 0 || xy->whitex > PNG_FP_1) return 1;
-   if (xy->whitey < 5 || xy->whitey > PNG_FP_1-xy->whitex) return 1;
-
    /* The reverse calculation is more difficult because the original tristimulus
     * value had 9 independent values (red,green,blue)x(X,Y,Z) however only 8
     * derived values were recorded in the cHRM chunk;
@@ -1451,16 +1469,16 @@ png_XYZ_from_xy(png_XYZ *XYZ, const png_xy *xy)
     * value of 2 indicates an internal error to the caller.
     */
    if (png_muldiv(&left, xy->greenx-xy->bluex, xy->redy - xy->bluey, 7) == 0)
-      return 2;
+      return 1;
    if (png_muldiv(&right, xy->greeny-xy->bluey, xy->redx - xy->bluex, 7) == 0)
-      return 2;
+      return 1;
    denominator = left - right;
 
    /* Now find the red numerator. */
    if (png_muldiv(&left, xy->greenx-xy->bluex, xy->whitey-xy->bluey, 7) == 0)
-      return 2;
+      return 1;
    if (png_muldiv(&right, xy->greeny-xy->bluey, xy->whitex-xy->bluex, 7) == 0)
-      return 2;
+      return 1;
 
    /* Overflow is possible here and it indicates an extreme set of PNG cHRM
     * chunk values.  This calculation actually returns the reciprocal of the
@@ -1473,9 +1491,9 @@ png_XYZ_from_xy(png_XYZ *XYZ, const png_xy *xy)
 
    /* Similarly for green_inverse: */
    if (png_muldiv(&left, xy->redy-xy->bluey, xy->whitex-xy->bluex, 7) == 0)
-      return 2;
+      return 1;
    if (png_muldiv(&right, xy->redx-xy->bluex, xy->whitey-xy->bluey, 7) == 0)
-      return 2;
+      return 1;
    if (png_muldiv(&green_inverse, xy->whitey, denominator, left-right) == 0 ||
        green_inverse <= xy->whitey)
       return 1;
@@ -1520,25 +1538,14 @@ png_XYZ_from_xy(png_XYZ *XYZ, const png_xy *xy)
 static int
 png_XYZ_normalize(png_XYZ *XYZ)
 {
-   png_int_32 Y;
+   png_int_32 Y, Ytemp;
 
-   if (XYZ->red_Y < 0 || XYZ->green_Y < 0 || XYZ->blue_Y < 0 ||
-      XYZ->red_X < 0 || XYZ->green_X < 0 || XYZ->blue_X < 0 ||
-      XYZ->red_Z < 0 || XYZ->green_Z < 0 || XYZ->blue_Z < 0)
+   /* Normalize by scaling so the sum of the end-point Y values is PNG_FP_1. */
+   Ytemp = XYZ->red_Y;
+   if (png_safe_add(&Ytemp, XYZ->green_Y, XYZ->blue_Y))
       return 1;
 
-   /* Normalize by scaling so the sum of the end-point Y values is PNG_FP_1.
-    * IMPLEMENTATION NOTE: ANSI requires signed overflow not to occur, therefore
-    * relying on addition of two positive values producing a negative one is not
-    * safe.
-    */
-   Y = XYZ->red_Y;
-   if (0x7fffffff - Y < XYZ->green_X)
-      return 1;
-   Y += XYZ->green_Y;
-   if (0x7fffffff - Y < XYZ->blue_X)
-      return 1;
-   Y += XYZ->blue_Y;
+   Y = Ytemp;
 
    if (Y != PNG_FP_1)
    {