Did you know that our version of requests was from 2015? Phwoar!
This bumps `jinja2`, `py`, `pyyaml`, `requests` and `urllib3`.
There's significant risk for regressions due to breaking changes,
though due to the dynamic nature of Python, they're tricky to track
down.
The potential breaking changes I'm expecting to potentially affect
us are:
* `requests@2.11.0`: No longer accepts non-strings as header values.
* `requests@2.16.0`: `requests.packages` namespace was removed due to
packages no longer being vendored. The namespace has been
incrementally restored over future releases, but it's unclear to
what degree.
* `requests@2.24.0`: Redirect resolution now only happens when
`allow_redirects` is `True`.
* `requests` version of `urllib3` was bumped from `1.13.1` to `1.26`,
unsure what repercussions that will have.
This also added dependencies to `flake8_requirements`. This is because
the vendored `importlib-metadata` update (and the new
`typing-extensions` dependency) is found by `pip`, but the
vendored `typing-extensions` is _not_ discovered (perhaps due to
details around "eggs" and the way we add vendored code to the sys.path).
So, `pip` tries to install it manually and fails when it finds no hash.
By re-compiling `flake8_requirements`, the necessary hashes are made
available.
Differential Revision: https://phabricator.services.mozilla.com/D108526
The maximum version check has caused more unnecessary failed
builds (for freshly up-to-date devs) than helpful guards against
tough-to-diagnose issues.
Differential Revision: https://phabricator.services.mozilla.com/D113675
irrespective of the optimization level. The code was previously using
the optimization level as a proxy for whether --enable-release is
enabled or not, which was invalidated by bug 1689284.
Differential Revision: https://phabricator.services.mozilla.com/D113730
It's a relatively minor one (two keyframes are leaked). I built latest
GTK3 from source and it wasn't reproducible locally. I think our GTK
version on automation might be too old and not have this commit:
10fa786dfa
Or something of that sort.
Differential Revision: https://phabricator.services.mozilla.com/D113948
Not all in-tree Rust libraries were using workspace-hack.
Additionally, some needed winapi features were missing from
workspace-hack's configuration.
Now, winapi is re-compiled less frequently on a full build.
Differential Revision: https://phabricator.services.mozilla.com/D113564
Sometimes there is a deadlock between user-requested ccov counters flush (through a signal) and flush happening when a process is dying.
So in order to avoid it, just remove the latter which is out of our control (user-requested flushes are protected using a cross mutex).
Differential Revision: https://phabricator.services.mozilla.com/D112632
Credit goes to dmajor.
Rust gets LLVM 12 - https://github.com/rust-lang/rust/pull/81451.
The PGO format changed between 11 and 12, and builds will have poor optimization if we mix versions in xLTO.
The above PR has landed on Rust nightly 1.52.
Until we move to 1.52 nightly we should have this landed.
In the meantime, for unblocking try builds, we can disable value-PGO in Rust.
Differential Revision: https://phabricator.services.mozilla.com/D113327
We should probably enable this later on when we do the aproapriate fixes in m-c and they
are also tagged in m-r but for now we should have this changed backed out of llvm locally.
Differential Revision: https://phabricator.services.mozilla.com/D113215
As an intermediate step to allow mach commands as standalone functions, the MachCommandBase
subclass instance that currently corresponds to self has to be made available as a separate
argument (named command_context).
Differential Revision: https://phabricator.services.mozilla.com/D109650
This both helps IDE integration and will help avoid build issues which would've
otherwise cropped up when adding new files to the directory for bug 1706374.
Differential Revision: https://phabricator.services.mozilla.com/D112764
Check if a user is a Mozilla employee by checking their
Bugzilla groups, or checking if their VCS email ends
with "@mozilla.com".
When a user is setting up a new build environment, telemetry
will be automatically enabled for them if they are an
employee. If they're not an employee, they'll be asked
if they want to opt in.
Differential Revision: https://phabricator.services.mozilla.com/D106315
Did you know that our version of requests was from 2015? Phwoar!
This bumps `jinja2`, `py`, `pyyaml`, `requests` and `urllib3`.
There's significant risk for regressions due to breaking changes,
though due to the dynamic nature of Python, they're tricky to track
down.
The potential breaking changes I'm expecting to potentially affect
us are:
* `requests@2.11.0`: No longer accepts non-strings as header values.
* `requests@2.16.0`: `requests.packages` namespace was removed due to
packages no longer being vendored. The namespace has been
incrementally restored over future releases, but it's unclear to
what degree.
* `requests@2.24.0`: Redirect resolution now only happens when
`allow_redirects` is `True`.
* `requests` version of `urllib3` was bumped from `1.13.1` to `1.26`,
unsure what repercussions that will have.
This also added dependencies to `flake8_requirements`. This is because
the vendored `importlib-metadata` update (and the new
`typing-extensions` dependency) is found by `pip`, but the
vendored `typing-extensions` is _not_ discovered (perhaps due to
details around "eggs" and the way we add vendored code to the sys.path).
So, `pip` tries to install it manually and fails when it finds no hash.
By re-compiling `flake8_requirements`, the necessary hashes are made
available.
Differential Revision: https://phabricator.services.mozilla.com/D108526
Glean is throwing an error if a string metric is `None`.
Since `get_cpu_brand()` can return `None`, only update telemetry if it's
set.
Depends on D110847
Differential Revision: https://phabricator.services.mozilla.com/D110858
Glean is throwing an error if a string metric is `None`.
Since `get_cpu_brand()` can return `None`, only update telemetry if it's
set.
Depends on D110847
Differential Revision: https://phabricator.services.mozilla.com/D110858
To avoid erroneously importing a package from the system, we remove
system site package.
This is another positive step towards further insulating the Firefox
build from the state of the environment.
This may cause failures if some mach commands have been sneakily
importing system packages instead of properly vendoring them.
Differential Revision: https://phabricator.services.mozilla.com/D110847
This file has been unchanged (other than license changes)
since the 1998 "free the lizard" commit.
It's not referenced from any other file.
It's time to retire, `genrc.sh`.
Differential Revision: https://phabricator.services.mozilla.com/D112374
As an intermediate step to allow mach commands as standalone functions, the MachCommandBase
subclass instance that currently corresponds to self has to be made available as a separate
argument (named command_context).
Differential Revision: https://phabricator.services.mozilla.com/D109650
1. Add task to get and build clang from the main branch.
2. Using clang main toolchain we build on a daily basis linux64 firefox, this tasks also automatically triggers the fetch and build of clang from main branch since we don't cache it.
Differential Revision: https://phabricator.services.mozilla.com/D111063
In bug 1174288 and related bugs we created a framework for generating
test certificates (and later, keys) from specifications at build time. This
turned out to take too long to run on each build, so this system was largely
left disabled (see all of the "# Temporarily disabled. See bug 1256495."
comments removed in this patch). This patch introduces a mach command
("generate-test-certs") that can generate test certificates and keys. The
expectation is that when a developer needs to add new such artifacts, they can
use this new command. Similarly, when the artifacts need to be updated (for
example, because they've expired), this command can regenerate them all at
once.
Differential Revision: https://phabricator.services.mozilla.com/D108869
Allow using the MOZ_KnownLive function to get around it.
Use case is the following: I have an std::function member variable, and I want
that member to be able to capture `this`.
Using a strong reference creates a cycle and thus would leak. I know `this` to
always outlive the member, so it is fine to use a weak capture there.
Differential Revision: https://phabricator.services.mozilla.com/D111850
Glean is throwing an error if a string metric is `None`.
Since `get_cpu_brand()` can return `None`, only update telemetry if it's
set.
Depends on D110847
Differential Revision: https://phabricator.services.mozilla.com/D110858
To avoid erroneously importing a package from the system, we remove
system site package.
This is another positive step towards further insulating the Firefox
build from the state of the environment.
This may cause failures if some mach commands have been sneakily
importing system packages instead of properly vendoring them.
Differential Revision: https://phabricator.services.mozilla.com/D110847
This also makes `MOZ_UPDATE_AGENT` depend on `MOZ_BACKGROUNDTASKS`.
There's no need to lint the `tools/update-programs` project, since
it's not a supported configuration. The next consumer can
rehabilitate it or remove it.
Differential Revision: https://phabricator.services.mozilla.com/D111313
If a custom `topsrcdir` is needed, it can be directly passed in.
I don't see how a separate `mozilla_dir` parameter is necessary - it
appears redundant to me.
Differential Revision: https://phabricator.services.mozilla.com/D110841
The build always uses them anyways, but configure tests don't, and some
of them fail as a consequence in some cases (example: when UBsan is
enabled, all AC_LANG_CPLUSPLUS + AC_CHECK_FUNCS tests fail). We also had
similar problems in the past, such as bug 1513605.
By adding the PIC flags to CFLAGS and CXXFLAGS, we ensure old-configure
tests use the flags as well.
While here, we also remove DSO_CFLAGS, which was always empty, and we
stop passing -fPIC to wasm compiles, because it has no effect there.
Differential Revision: https://phabricator.services.mozilla.com/D108560
The build always uses them anyways, but configure tests don't, and some
of them fail as a consequence in some cases (example: when UBsan is
enabled, all AC_LANG_CPLUSPLUS + AC_CHECK_FUNCS tests fail). We also had
similar problems in the past, such as bug 1513605.
By adding the PIC flags to CFLAGS and CXXFLAGS, we ensure old-configure
tests use the flags as well.
While here, we also remove DSO_CFLAGS, which was always empty, and we
stop passing -fPIC to wasm compiles, because it has no effect there.
Differential Revision: https://phabricator.services.mozilla.com/D108560
At the same time, remove the global define, which isn't necessary and
triggers some rebuilds. Nothing is using MOZ_UPDATE_AGENT right now
so it's safe to enable more broadly, and it will be used to gate new
background update work.
Differential Revision: https://phabricator.services.mozilla.com/D102786
Our bot was expecting to have tasks that were succeeding, right now we no longer need that
and in a case of a build failure, or in a case of the presence of a checker that asserts itself
as error and not warning we should mark the task red by failing the job.
Differential Revision: https://phabricator.services.mozilla.com/D107656
At the same time, remove the global define, which isn't necessary and
triggers some rebuilds. Nothing is using MOZ_UPDATE_AGENT right now
so it's safe to enable more broadly, and it will be used to gate new
background update work.
Differential Revision: https://phabricator.services.mozilla.com/D102786
We already know the project name, we should unbundle it
from the project path.
This change is possible due to the work in 1664083,
and puts us back to the state before 1255185.
Differential Revision: https://phabricator.services.mozilla.com/D106293
Just adding the toolchain tasks for now. They are not yet used by anything, but bug 1693288 might be interested eventually.
As with earlier releases, clang-tidy and mingw builds are not included here, they'll be separate patches.
Differential Revision: https://phabricator.services.mozilla.com/D106554
Upstream wptserve just switched to Python 3 only. That's fine for
web-platform-tests, but it turns out that some marionette harness
tests are also using wptserve and are still on Python 2.
Since fixing marionette harness turns out to be non-trivial and this
blocks other wpt work, this patch does the following:
* Temporarily vendors the last wptserve revision that works with
Python 2 in to testing/web-platform/mozilla/tests/tools/wptserve_py2
* Configures the mach virtualenv to use that copy for Python 2 modules
only.
* Configures the test packaging system to also put that copy in the
common tests zip. Requirements files are updated to use either the
Python 2 version or the Pyhton 3 version as required.
Differential Revision: https://phabricator.services.mozilla.com/D106764
Spidermonkey package builds don't have MOZ_AUTOMATION set (for good
reasons), which means they don't automatically get the sysroot through
MOZ_FETCHES_DIR. We need a way for that particular build to set its
sysroot.
Because the setup for sysroot is more elaborated than adding --sysroot
to C{,XX}FLAGS, it's more convenient to have an option for this.
And while at it, we might as well make that a full-fledged option,
although, we only make it available when targetting Linux (although it
could be useful for other OSes, but there's overlap with other options
on Android, and Mac, so for now, Linux-only will do).
This also allows to pass --without-sysroot along with
--enable-bootstrap, where the bootstrapped sysroot is not used.
Differential Revision: https://phabricator.services.mozilla.com/D106352
