* Tue Jun 13 2017 Kevin Kofler <Kevin@tigcc.ticalc.org> - 5.9.0-2
- arm-fpu-fix patch: Also build the host tools (i.e., GN) with the correct FPU
This ensures GN does not get built with NEON on 32-bit armv7hl. The ARM
builders don't necessarily support NEON. (The Fedora one apparently
does, but the RPM Fusion one apparently does not.)
* Add the dependency on the SSE2 build to a target QtWebEngine actually
uses.
* Also handle v8_build_shared in the V8 inspector, so the symbols
actually get exported.
- Backport upstream patch to fix native builds on ARM and aarch64
(QTBUG-61128)
- Drop "Backport patch to fix FTBFS with GCC on aarch64 from upstream
Chromium"
- Drop "Delete all "toolprefix = " lines from
build/toolchain/linux/BUILD.gn"
Runtime detection would be nice to have, but for now this is the same
state as in 5.8. NEON is also not runtime-detected there. The x86_64 and
aarch64 builds will get the SIMD optimizations, the i686 and armv7hl
ones won't for now.
* add missing configs for the blink_x86_sse2 source set in
src/3rdparty/chromium/third_party/WebKit/Source/platform/BUILD.gn
* in src/core/core_module.pro, when processing NINJA_SOLIBS, ignore the
dummy convert_dict shared library used only to get a .pri file
Instead of relying on base/cpu.h, copy its implementation of __cpuid for
GCC/Clang (from base/cpu.cc) and use the original SSE2 detection code
from ANGLE using __cpuid.
- Build the SSE2 code in a separate compilation unit on i686.
- Make supportsSSE2() return always true rather than always false on
x86_64.
- Make supportsSSE2() call the detection code from base/cpu.h instead of
always returning false on i686.
* Sat Jun 10 2017 Kevin Kofler <Kevin@tigcc.ticalc.org> - 5.9.0-1
- Update to 5.9.0
- Update version numbers of bundled stuff
- Use bundled libsrtp and protobuf, Chromium dropped unbundling support for them
- Add missing Provides: bundled(hunspell) for the spellchecking added in 5.8
- Rebase linux-pri, no-neon, system-icu-utf, no-sse2, arm-fpu-fix,
openmax-dl-neon and webrtc-neon-detect patches (port to GN)
- Sync system-nspr-prtime patch with Debian (they ported it to GN)
- Rebase fix-extractcflag patch
- Restore NEON runtime detection in Skia, drop old skia-neon patch (rewritten)
- Drop webrtc-neon, v8-gcc7, pdfium-gcc7, wtf-gcc7, fix-open-in-new-tab and
fix-dead-keys patches, fixed upstream
- Update system libvpx/libwebp version requirements (libvpx now F25+ only)
- Drop the flag hacks (-g1 -fno-delete-null-pointer-checks) that are fixed
upstream, force -g2 on x86_64 instead
There are no longer 'ffmpeg_branding == "Chromium"' conditionals, we
need to key on the absence of 'ffmpeg_branding == "Chrome"' or
'ffmpeg_branding == "ChromeOS"' conditionals instead.
This reverts commit 7816e88f62.
I already told you last time (5.8) that this way of maintaining patches
(importing them all into git and reexporting them from there each time)
is not how I work (I want only those patches rebased and renamed that
actually need rebasing) and that you are only causing me more work.
Please keep your hands off my package from now on.
This reverts commit 024052080f.
I already told you last time (5.8) that this way of maintaining patches
(importing them all into git and reexporting them from there each time)
is not how I work (I want only those patches rebased and renamed that
actually need rebasing) and that you are only causing me more work.
Please keep your hands off my package from now on.
This reverts commit ee046f9af5.
I already told you last time (5.8) that this way of maintaining patches
(importing them all into git and reexporting them from there each time)
is not how I work (I want only those patches rebased and renamed that
actually need rebasing) and that you are only causing me more work.
Please keep your hands off my package from now on.
