Linux 7.2-rc1

28 June 19:01 - Linux Kernel - Linus Torvalds

[autofit] Derive fill orientation without a separate outline pass

28 June 09:20 - FreeType - Behdad Esfahbod

`af_glyph_hints_reload` called `FT_Outline_Get_Orientation` once per glyph solely to choose the major-direction sign.

* meson.build, .gitlab-ci.yml: Require Meson 0.60.0 or newer

27 June 04:44 - FreeType - TheLastRar

Version 0.

nss: Remove --enable-static-nss configure option

25 June 13:20 - glibc - Michael Ford

Since the NSS reorganization in glibc 2.33 the --enable-static-nss option no longer changes the build: the files and dns services are built into libc unconditionally, and there is no longer any mechanism to bake the remaining services into libc.a.

ppc64le: Restore optimized memchr for power10 [BZ #34300]

25 June 05:21 - glibc - Sachin Monga

Restore the POWER10 memchr implementation reverted in commit a7877bb6685300f159fa095c9f50b22b112cddb8 [BZ #33059], addressing the actual cause: The non-volatile vector register v20 used as the zero vector (M_VREG_ZERO) is replaced with the volatile register v17.

swscale/ops: switch from AVRational to AVRational64

24 June 23:20 - FFmpeg - Niklas Haas

This has two immediate consequences: 1.

avutil/int128: add header for 128-bit integers

24 June 23:20 - FFmpeg - Niklas Haas

These can be implemented efficiently on most modern 64-bit compilers.

swscale/rational64: add 64-bit rational type

24 June 23:20 - FFmpeg - Niklas Haas

This is needed by the ops code, to represent intermediate values for 32-bit formats, which can exceed the value range of int32_t (especially for intermediate products). I copied the math almost 1:1 from rational.c, but adapted to use the 128-bit integer wrappers defined by int128.h.

avcodec/nvenc: remove support for SDK versions older than 11.1

24 June 18:51 - FFmpeg - Timo Rothenpieler

avfilter/cuda: support P012/P212 and MSB 4:4:4 in scale/transpose/thumbnail

24 June 12:18 - FFmpeg - Diego de Souza

NVDEC and CUVID now output AV_PIX_FMT_P012 (12-bit 4:2:0), AV_PIX_FMT_P212 (12-bit 4:2:2) and AV_PIX_FMT_YUV444P10MSB / AV_PIX_FMT_YUV444P12MSB (10/12-bit 4:4:4) for high-bit-depth content, but these CUDA filters rejected the formats in their supported-format lists, breaking pipelines such as "-hwaccel cuda ...

avutil/hwcontext_cuda: add P012 and P212 to supported formats

23 June 23:58 - FFmpeg - Diego de Souza

NVDEC and CUVID decode 12-bit 4:2:0 content to AV_PIX_FMT_P012 and 12-bit 4:2:2 to AV_PIX_FMT_P212, but these formats were missing from the CUDA frames context supported format list.

avfilter/vf_scale_cuda: add generic 1D filter kernel

23 June 19:22 - FFmpeg - Niklas Haas

This can be useful for any sort of separable filtering with arbitrary weights.

avfilter/vf_scale_cuda: add `use_filters` option

23 June 19:22 - FFmpeg - Niklas Haas

This may be faster or slower than the existing specialized kernels, so I opted not to prefer it by default.

avformat/http: make short-seek logic more robust

23 June 18:33 - FFmpeg - Niklas Haas

This avoids an underflow if short_seek is negative (which can happen if e.g.

AArch64: Update SHARED-FILES with files from Arm Optimized-Routines

23 June 17:50 - glibc - Pierre Blanchard

Adds lists for math.

swscale/ops: keep track of copied/cleared components

23 June 11:48 - FFmpeg - Niklas Haas

These represent components which have not (yet) been modified from their input values (i.e.

swscale/ops_dispatch: add option to link subpass outputs together

23 June 11:48 - FFmpeg - Niklas Haas

Not needed currently but will be used for parallel splits.

swscale/ops_dispatch: add option to split const/copied subpasses

23 June 11:48 - FFmpeg - Niklas Haas

This already helps performance as-is, but will help performance massively once we add the ability for the memcpy backend to do a refcopy instead of an actual copy.

swscale/tests/sws_ops: split passes when printing ops lists

23 June 11:48 - FFmpeg - Niklas Haas

This affects a large number of conversions across the board, either: 1.