We should only observe for update errors while we are expecting
a successful update.
MozReview-Commit-ID: 3grGhmxqhIX
--HG--
extra : rebase_source : d099b83560ac5ec15b18fb69177368a645b63952
It turns out sometimes (in the LTO+CFI case at least) Assertions.h
will not be present in the opt build, presumably because it was optimized
out.
MozReview-Commit-ID: GB3GIoSdIUK
mach android-emulator currently supports 6 different avds; I am struggling to maintain
that many configurations. I don't see a lot of value in keeping both 6.0 and 7.0,
and Android 6.0 is not as popular as 7.0. Let's remove 6.0, encouraging 7.0 as an
alternative; same for x86-6.0 -> x86-7.0.
1. Updated hgrepo to work with mozilla-beta, mozilla-esr60 and project branches (just in case)
2. Presquashed commits, so we only submit one.
3. Replaced 'which' with 'command -v' to avoid future shellcheck issues.
Differential Revision: https://phabricator.services.mozilla.com/D1582
This adds 'CorruptionCanaryForStatics', which as the name implies is suitable
for use in objects that are statically declared. It has a trivial destructor
which allows us to avoid the need for static constructors.
--HG--
extra : amend_source : 27f8eff9ead21fde9f5f5d17f16c322d2c995a27
This changes two config options:
pytest_classes = PyTest # only classes that start with 'PyTest' will be considered tests (previously this was Test)
xfail_strict = true # tests marked as xfail will cause pytest to return non-zero if they unexpectedly pass
MozReview-Commit-ID: DCWoDFbe6Mk
--HG--
extra : rebase_source : 9aa806e035d62d51bb338708396851c40f55ee00
When the Windows OS shuts down, we use a synchronous shutdown mechanism,
this exercises session save and restore in a unique way.
MozReview-Commit-ID: 6sCa3E2wmLY
--HG--
extra : rebase_source : 05014c26faa932165b03f922a63ec9576462bc67
This version of the blocklist should be functionally comparable to the mozglue
based blocklist, except:
* We hook NtMapViewOfSection instead of LdrLoadDll: The former allows us to
easily obtain the module file name being used for the load. The latter
requires us to essentially emulate the loader's path searching, which is
a perf hit, potentially a correctness issue, and more work to do given the
limited native NT API set.
* Since the paths in native NT land are all unicode, and since this code is
critical to startup performance, this version of the blocklist uses unicode
strings instead of ASCII strings. My thoughts here are that we don't want to
be wasting time on every DLL load doing ASCII-to-unicode conversion every
time we want to do a blocklist string comparison.
* I am completely aware that this leaves us in a bizarre situation where we have
two copies of the blocklist in our binaries: one unicode version in firefox.exe,
and one ASCII version in mozglue.dll. Once we (hopefully) move to using the
launcher process by default, the ASCII copy can go away. In the meantime, we
need to be able to use either one depending on how Firefox was started. I am
happy to make the Native NT blocklist Nightly-only to assuage these concerns.
* This allows us to use a single blocklist definition in multiple places.
* This patch also adds support for a new initialization flag that, when enabled,
disables the DLL Blocking part of the mozglue blocklist but leaves the
profiling and stackwalking suppression bits intact.
Leave support for multil10n uses for now.
for "Cleanup l10n mozharness config files." (Batch 2)
Differential Revision: https://phabricator.services.mozilla.com/D1561
--HG--
extra : source : bb28c2d9d5f51f6e7eb797b892247948f1f7021b