Source history does not give any good clues about why this line was added in the first place. In any case, LLVM trunk currently has build bustage when threads are disabled. We could work around the bustage and/or wait for a fix, but it seems like threads are a good thing to have in general nowadays. Maybe this could help with LTO build times.
Differential Revision: https://phabricator.services.mozilla.com/D82447
The logic was apparently initially copied from the compositor thread which too got migrated to nsThread in bug 1634253, so we re-use that same logic again.
This allows to remove all use of the Google's MessageLoop.
Differential Revision: https://phabricator.services.mozilla.com/D81974
There's a small race that can happen when the remote decoder gets shutdown during xpcom shutdown; that would cause GetCurrentSerialEventTarget to return null. Leading to an assertion failure in ActorLifecycleProxy thread-safety check when PRemoteDecoderManagerParent gets destroyed.
So we use a background taskqueue instead and cleanup a bit the threading code in there allowed thanks to the TaskQueue ability to not require an explicit shutdown.
Differential Revision: https://phabricator.services.mozilla.com/D81287
Because jar_maker is not in the libs tier, we also rename the libs-%
targets for l10n repacks to l10n-%, which make it clearer what they are
for.
And because multilocale.txt is both a GeneratedFile and a file that is
generated (and installed) via manual build rules, keeping it in the misc
target actually breaks building in toolkit/locales during l10n repacks,
so move it to libs for now.
Differential Revision: https://phabricator.services.mozilla.com/D81766
We get a lot of bugs about building Firefox for Android not working, and there's been ongoing confusion about whether this is something that does work or who has ever worked. The warning should hopefully clarify the current situation for people.
Differential Revision: https://phabricator.services.mozilla.com/D81585
There's a small race that can happen when the remote decoder gets shutdown during xpcom shutdown; that would cause GetCurrentSerialEventTarget to return null. Leading to an assertion failure in ActorLifecycleProxy thread-safety check when PRemoteDecoderManagerParent gets destroyed.
So we use a background taskqueue instead and cleanup a bit the threading code in there allowed thanks to the TaskQueue ability to not require an explicit shutdown.
Differential Revision: https://phabricator.services.mozilla.com/D81287
clang 11 will stop using full paths in PGO dependentlib declarations, so we need to tell the linker where to find clang_rt.profile-x86_64.lib.
This is roughly a revert of bug 1621552.
Differential Revision: https://phabricator.services.mozilla.com/D81244
Some of the testing info is from the Oxidation wiki, and the logging info is
largely from a dev-platform email by Valentin. The other parts I wrote from
scratch.
The commit also makes some small improvements to the Rust build docs.
Differential Revision: https://phabricator.services.mozilla.com/D81017
And add suppressions for the new errors that this adds to the valgrind
run. It's not clear if it's a legitimate thing that LLVM does that
valgrind doesn't cope with, like many others, but the fact is running
valgrind on a PGO/LTO build doesn't show these errors, so we're not
actually shipping them anyways (but does show _different_ errors that we
don't see on the build the valgrind jobs do)
Differential Revision: https://phabricator.services.mozilla.com/D81016
In clang-11 the flag's behavior became the default, and in fact the flag is no longer accepted so we get an unrecognized option error.
Differential Revision: https://phabricator.services.mozilla.com/D81124
The added material comes from the Oxidation wiki's FAQ and the "Building
Firefox with Rust code" MDN page:
https://wiki.developer.mozilla.org/en-US/docs/Mozilla/Firefox/Building_Firefox_with_Rust_code
Once this commit lands, I will (a) remove the wiki FAQ entries, and (b) replace
the contents of that MDN page with a link to these build docs. There's no point
having overlapping docs, and in-tree is the best place for this information.
This commit also makes the capitalization of sub-headings more consistent.
Differential Revision: https://phabricator.services.mozilla.com/D79956
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
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
When linking a weak symbol in an object against a library where the
symbol is provided with a version, the final binary gets 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
In adcd026838, conversions from `StringRef` to `std::string` became `explicit`.
The NFC fix would be to wrap the conversions in `std::string(...)` but since these are all short-lived uses for things like string compares, I've refactored a bit to keep them in native form.
Differential Revision: https://phabricator.services.mozilla.com/D79895
Sometime during clang 11 (I don't have the exact commit handy), the display of nested templates changed from `Foo<Bar<X> >` to `Foo<Bar<X>>`.
Differential Revision: https://phabricator.services.mozilla.com/D79894
When linking a weak symbol in an object against a library where the
symbol is provided with a version, the final binary gets 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
I wrote this patch to address two problems:
1. if I do `mach run` from a directory other than $topsrcdir, $objdir, or
$objdir/dist/bin, then .gdbinit will not be loaded.
2. Debugging the firefox binary will never load the JS prettyprinters in any
case.
I believe this patch fixes other problems as well, such as .gdbinit_python not being found, and the gdbpp pretty-printers not getting loaded in various situations.
The main changes of this patch are:
1. Move .gdbinit into build/ (and $objdir/build/) to delay it from getting
loaded until the search path is configured.
2. Move libxul.so-gdb.py into the correct directory.
3. Use either libxul.so-gdb.py or js-gdb.py to configure the correct
search path then load .gdbinit, and have .gdbinit load all of the
pretty-printers (Gecko and JS).
4. Use a single preprocessed file to configure the source directory. Use
relative paths within the objdir for everything else.
Differential Revision: https://phabricator.services.mozilla.com/D77589
This brings the `android-build` image very close to other build
images, paving the way for it to be folded in completely. It also
makes us more resilient in the face of upstream service interruptions.
Differential Revision: https://phabricator.services.mozilla.com/D78945
I wrote this patch to address two problems:
1. if I do `mach run` from a directory other than $topsrcdir, $objdir, or
$objdir/dist/bin, then .gdbinit will not be loaded.
2. Debugging the firefox binary will never load the JS prettyprinters in any
case.
I believe this patch fixes other problems as well, such as .gdbinit_python not being found, and the gdbpp pretty-printers not getting loaded in various situations.
The main changes of this patch are:
1. Move .gdbinit into build/ (and $objdir/build/) to delay it from getting
loaded until the search path is configured.
2. Move libxul.so-gdb.py into the correct directory.
3. Use either libxul.so-gdb.py or js-gdb.py to configure the correct
search path then load .gdbinit, and have .gdbinit load all of the
pretty-printers (Gecko and JS).
4. Use a single preprocessed file to configure the source directory. Use
relative paths within the objdir for everything else.
Differential Revision: https://phabricator.services.mozilla.com/D77589
To avoid sending identifying information, common absolute paths are patched with placeholder values. For example, devs
may place their Firefox repository within their home dir, so absolute paths are doctored to be prefixed with
"<topsrcdir"> instead.
Additionally, any paths including the user's home directory are patched to instead be a relate path from "~".
Differential Revision: https://phabricator.services.mozilla.com/D78962
To avoid sending identifying information, common absolute paths are patched with placeholder values. For example, devs
may place their Firefox repository within their home dir, so absolute paths are doctored to be prefixed with
"<topsrcdir"> instead.
Additionally, any paths including the user's home directory are patched to instead be a relate path from "~".
Differential Revision: https://phabricator.services.mozilla.com/D78962
Rather than always printing instructions at the end of the bootstrap phase, we will now create a mozconfig
file if one doesn't exist and there's configuration to be written.
Differential Revision: https://phabricator.services.mozilla.com/D78417
For standalone JS builds there are three possibilities: no-NSPR (the default),
build NSPR, or use system NSPR.
This patch makes no-NSPR the default again when not overridden by --enable-nspr-build
or --with-system-nspr. This was likely regressed by bug 1641775.
Differential Revision: https://phabricator.services.mozilla.com/D78001
clang_getAddressSpace was found as a clang 5.0+ API by comparing the "tools/libclang/libclang.exports" files in the clang source code.
"clang_getAddressSpace" exists for 5.0.0 and 10.0.0, but not for 4.0.1.
Differential Revision: https://phabricator.services.mozilla.com/D78374
When not using --with-android-ndk, NDK root is `~/.mozbuild/android-ndk-r20`.
When building this NDK path, since we use `os.path.join`, so it may include "\"
as file separator on Windows host. Then some options such as `-gcc-toolchain`
become invalid path (ex. `C:/users/makoto.mozbuildandroid-ndk-r20/`..).
So we should replace file separator with "/".
Differential Revision: https://phabricator.services.mozilla.com/D77820
When running |mach python| with python 3, it was failing when importing "enum", since it would get the python 2 backport. By making these packages only
available to python 2, we allow python 3 to import standard library packages properly.
Differential Revision: https://phabricator.services.mozilla.com/D77960
Version of NSS >= 3.27 have a pkg-config file. We're now requiring 3.53,
so while moving, just use pkg-config, which is simpler.
The old-configure check that rejected some untested platforms for
in-tree NSS is actually rejecting none: the accepted platforms cover all
the supported ones, so we remove that check.
And because building with system NSS without system NSPR doesn't make
sense, imply the latter when the former is used.
Differential Revision: https://phabricator.services.mozilla.com/D77428
At the same time, because it's now simpler to do so, set the right data
file name for big-endians, even though we don't have or produce it
(bug #1264836). Also remove USE_ICU, which is redundant with
JS_HAS_INTL_API, and actively break the build at configure time when
using --without-intl-api with Firefox because this hasn't actually
worked for close to 3 years (since bug 1402048).
Differential Revision: https://phabricator.services.mozilla.com/D77416
Versions of NSPR >= 4.10 come with a pkg-config file. We currently
depend on 4.9.2 for spidermonkey, but much more recent versions for
Firefox. 4.10 is less than a year newer than 4.9.2, and 4.10 is 7 years
old, so bumping the requirement to 4.10 is not really a big deal.
With the use of pkg-config, --with-nspr-cflags and --with-nspr-libs are
not needed.
None of the AC_TRY_COMPILE tests were any useful because
PR_STATIC_ASSERT and PR_UINT64 have been when we look for them since
4.8.6 and 4.9 respectively.
Differential Revision: https://phabricator.services.mozilla.com/D77412
As all versions of zlib >= 1.2.3.1 have a pkg-config file, and 1.2.3.1
is close to 14 years old, let's drop 1.2.3 and just use pkg-config, which
simplifies what we need to do dramatically.
Differential Revision: https://phabricator.services.mozilla.com/D77404
l10n-merge creates a full merge dir for a while now, let's
simplify the build logic to only read from that directory
during repacks and langpacks.
Differential Revision: https://phabricator.services.mozilla.com/D77023