When linking a weak symbol in an object against a library where the
symbol is provided with a version, the final binary get a weak versioned
symbol reference.
It turns out weak versioned symbols still make the dynamic linker need
the symbol version, even if all symbols needed with that version are
weak.
Practically speaking, that means with bug 1634204, we now end up with
a weak versioned symbol reference to __cxa_thread_atexit_impl with
version GLIBC_2.18, and glibcs without the symbol can't fulfil that
version, even though the weak symbol is the only thing we need from that
version.
This means the check_binary changes in bug 1634204 are too
relaxed, so we revert them (although we keep the easier to read
conditions in check_dep_versions).
We also introduce a hack in stdc++compat.cpp (although it's not
technically entirely about libstdc++ compat) so that we avoid the weak
symbol reference while keeping the intended baseline for libstdc++ and
glibc.
Differential Revision: https://phabricator.services.mozilla.com/D79773
With containerless scrolling we generally always want this clip on the layer,
since there is no container. However on Android apparently we don't want this,
because it seems to break the setVerticalClipping API used by the dynamic
toolbar.
Differential Revision: https://phabricator.services.mozilla.com/D80054
Content process won't try to send IPC to parent process to update cursor if the
cursor isn't changed, and assume parent process would update cursor properly as
soon as the cursor re-enter to a remote target.
But BrowserParent cached the cursor information for a remote target only when the
remote target is allowed to update the cursor. It could possible that parent use
a stale cursor for a remote target and stuck in that state untill content
process tries to update cursor again.
Differential Revision: https://phabricator.services.mozilla.com/D79985
Content process could also contain a remote target after fission, we need to do
same check in content process, like what we do in parent process for e10s.
Depends on D79441
Differential Revision: https://phabricator.services.mozilla.com/D79917
This patch adds a `--test-date` option to mozperftest that lets us specify the date that we are testing on. This lets us trigger multi-commit tests on past dates, and also ensures that retriggers of the multi-commit tests use the same day that they were pushed on (giving us reproducibility).
Differential Revision: https://phabricator.services.mozilla.com/D80173
We've had nsNativeBasicTheme enabled since 75, and all reported issues
were fixed real soon (and I haven't heard of any of them recently).
Given the non-native theme is likely changing in the future, I'd rather
not maintain three themes for Android :)
Differential Revision: https://phabricator.services.mozilla.com/D80105
They were necessary some time ago, but:
- there is no js/xpconnect/src/export target anymore,
- accessibility/xpcom/export had a dependency on xpcom/xpidl/export for
the ply cache files, but they're not used as of bug 1633156.
- the recursive make backend was adding automatic dependencies on
xpcom/xpidl/export for the ply cache files.
Differential Revision: https://phabricator.services.mozilla.com/D80150
When linking a weak symbol in an object against a library where the
symbol is provided with a version, the final binary get a weak versioned
symbol reference.
It turns out weak versioned symbols still make the dynamic linker need
the symbol version, even if all symbols needed with that version are
weak.
Practically speaking, that means with bug 1634204, we now end up with
a weak versioned symbol reference to __cxa_thread_atexit_impl with
version GLIBC_2.18, and glibcs without the symbol can't fulfil that
version, even though the weak symbol is the only thing we need from that
version.
This means the check_binary changes in bug 1634204 are too
relaxed, so we revert them (although we keep the easier to read
conditions in check_dep_versions).
We also introduce a hack in stdc++compat.cpp (although it's not
technically entirely about libstdc++ compat) so that we avoid the weak
symbol reference while keeping the intended baseline for libstdc++ and
glibc.
Differential Revision: https://phabricator.services.mozilla.com/D79773
This removes a dependency on `pymake`, which is Python 2-only, and thoroughly unnecessary since we just use it to find assignments of the form `OBJECTS = ...`. We can replicate this logic by just isolating lines that begin with that literal string, and everything else can stay the same. This is definitionally more brittle than actually using a parser, but it works fine for now, and the original implementation wasn't significantly better (it didn't handle any form of dynamism, anything more complicated than a single unconditional assignment with a space-separated list of literal strings representing outputs, etc.)
Differential Revision: https://phabricator.services.mozilla.com/D79896
Sentry needs to be able to send data in a minimal environment, and intelligently determining
the "topobjdir" (without leaning on Mach) is tough.
Instead, since the "topobjdir" usually falls under the "topsrcdir", just leaning on <topsrcdir>
patching should be sufficient.
Differential Revision: https://phabricator.services.mozilla.com/D80122
Chrome will not accept sdp with ssrc-group:FID on recvonly msections, causing
calls to fail to be established between Firefox and Chrome when only a
microphone is present on the Firefox side.
Differential Revision: https://phabricator.services.mozilla.com/D80034
Byron Campen [:bwc]
Mike Hommey
Kartikaya Gupta
Edgar Chen
Mike Conley
Harald Kirschner
Gregory Mierzwinski
Daniel Lu
Andrew Creskey
Emilio Cobos Álvarez
Mitchell Hentges
Dorel Luca
Ricky Stewart
Marco Castelluccio
sefeng
Sam Foster
Sylvestre Ledru
Agi Sferro
Kris Taeleman
Nihanth Subramanya
Robert Helmer
Dan Minor
Jeff Muizelaar