A new pixman release 0.31.2 is now available. This is a release candidate for a stable 0.32.0 release.
This release contains performance improvements to bilinear scaling, and a large number of bug fixes. Please test, and report bugs to the mailing list at [email protected]. Søren tar.gz: http://cairographics.org/snapshots/pixman-0.31.2.tar.gz http://xorg.freedesktop.org/archive/individual/lib/pixman-0.31.2.tar.gz tar.bz2: http://xorg.freedesktop.org/archive/individual/lib/pixman-0.31.2.tar.bz2 Hashes: MD5: aced5dd2a4651c2bcd5c8c98bd7a64a9 pixman-0.31.2.tar.gz MD5: b347a6a44e5d3e7fdba644877250ce44 pixman-0.31.2.tar.bz2 SHA1: 0246796a1220a52e8e98110dd1460b64e7f2476c pixman-0.31.2.tar.gz SHA1: 5232cd11c178e32e8cc63f2678f777671470f74e pixman-0.31.2.tar.bz2 GPG signature: http://cairographics.org/snapshots/pixman-0.31.2.tar.gz.sha1.asc (signed by Søren Sandmann Pedersen <[email protected]>) Git: git://git.freedesktop.org/git/pixman tag: pixman-0.31.2 Log: Alexander Troosh (1): Require GTK+ version >= 2.16 Andrea Canciani (6): test: Fix build on MSVC Fix build on MSVC Fix building of "other" programs on MSVC Fix build of check-formats on MSVC Add support for SSSE3 to the MSVC build system Fix thread-test on non-OpenMP systems Brad Smith (1): Fix pixman build with older GCC releases Markos Chandras (1): Use AC_LINK_IFELSE to check if the Loongson MMI code can link Matt Turner (2): mmx: Document implementation(s) of pix_multiply(). iwmmxt: Disallow if gcc version is < 4.8. Matthieu Herrb (1): configure.ac: Don't use '+=' since it's not POSIX Ritesh Khadgaray (1): pixman_trapezoid_valid(): Fix underflow when bottom is close to MIN_INT Siarhei Siamashka (7): test: fix matrix-test on big endian systems test: safeguard the scaling-bench test against COW sse2: faster bilinear scaling (pack 4 pixels to write with MOVDQA) vmx: align destination to fix valgrind invalid memory writes vmx: there is no need to handle unaligned destination anymore sse2: bilinear fast path for src_x888_8888 configure.ac: check and use -Wdeclaration-after-statement GCC option Søren Sandmann Pedersen (48): Post-release version bump to 0.31.1 Make use of image flag in noop iterators Add ITER_SRC and ITER_DEST iter flags Add ITER_WIDE iter flag Always set the FAST_PATH_NO_ALPHA_MAP flag for non-BITS images noop: Keep information about iterators in an array of pixman_iter_info_t sse2: Replace the fetcher_info_t table with a pixman_iter_info_t table mmx: Replace the fetcher_info_t table with a pixman_iter_info_t table fast: Replace the fetcher_info_t table with a pixman_iter_info_t table general: Store the iter initializer in a one-entry pixman_iter_info_t tab Add _pixman_implementation_iter_init() and use instead of _src/_dest_init Delete the old src/dest_iter_init() functions Consolidate all the iter_init_bits_stride functions RELEASING: Add note about changing the topic of the #cairo IRC channel test/scaling-bench.c: New benchmark for bilinear scaling Add support for iter finalizers Add an iterator that can fetch bilinearly scaled images fast_bilinear_cover_init: Don't install a finalizer on the error path Make the first argument to scanline fetchers have type bits_image_t * Drop support for 8-bit precision in bilinear filtering general: Ensure that iter buffers are aligned to 16 bytes Add empty SSSE3 implementation ssse3: Add iterator for separable bilinear scaling demos/scale.ui: Allow subsample_bits to be 0 pixman-filter.c: Use 65536, not 65535, for fixed point conversion {scaling,affine,composite-traps}-test: Use compute_crc32_for_image() test: Share the image printing code test: Test negative strides fast, ssse3: Simplify logic to fetch lines in the bilinear iterators Move bits_image_fetch_bilinear_no_repeat_8888 into pixman-fast-path.c Move generated affine fetchers into pixman-fast-path.c utils.c: Make print_image actually cope with negative strides utils.c: Make image_endian_swap() deal with negative strides blitters-test: Remove unused variable Rename HAVE_PTHREAD_SETSPECIFIC to HAVE_PTHREADS test: Add new thread-test program fast: Swap image and iter flags in generated fast paths pixman-combine32.c: Formatting fixes pixman-combine32: Improve documentation for blend mode operators pixman-combine32: Rename a number of variable from sa/sca to as/s pixman-combine32: Make Color Dodge routine follow the math more closely pixman-combine32.c: Make Color Burn routine follow the math more closely scale.c: Use int instead of kernel_t for values in named_int_t configure.ac: Add check for pmulhuw assembly utils.c: On x86-32 unalign the stack before calling test_function pixman-glyph.c: Add __force_align_arg_pointer to composite functions test/trap-crasher.c: Add trapezoid that demonstrates a crash Pre-release version bump to 0.31.2 [email protected] (1): Fix broken build when HAVE_CONFIG_H is undefined, e.g. on Win32. _______________________________________________ Pixman mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/pixman
