Gecko always adds .2em space above and below stretchy operators but this
is not part of MathML Core and is causing several WPT failures. The
MATH table does not have any parameter for that so it's likely some
tweak that was added in the early days of MathML that may no longer be
necessary. This patch tentatively removes this feature by placing it
under a disabled-by-default flag.
Differential Revision: https://phabricator.services.mozilla.com/D207014
In release builds we want to collect telemetry which is 1) useful and 2) is
likely to be different to that of the beta or nightly populations.
This patch adds collection of the the following telemetry in release builds by
default:
+ GC_ZONE_COUNT
+ GC_ZONES_COLLECTED
+ MEMORY_JS_GC_HEAP
+ GC_IN_PROGRESS_MS
+ GC_MARK_MS
+ GC_SWEEP_MS
+ GC_COMPACT_MS
+ GC_SLICE_MS
+ GC_MMU_50
+ GC_MINOR_US
The following are removed from collection in release builds by default, because
they are generally determined by scheduling code and not by the release
population:
- GC_BUDGET_MS_2
- GC_TIME_BETWEEN_S
The following are removed because they're not useful enough to justify:
- GC_ANIMATION_MS
- GC_PRETENURE_COUNT_2
- GC_TENURED_SURVIVAL_RATE
- GC_EFFECTIVENESS
The following are removed because they are part of more general telemetry and
not useful enough on their own:
- GC_MARK_ROOTS_US
- GC_MARK_GRAY_MS_2
- GC_MARK_WEAK_MS
Differential Revision: https://phabricator.services.mozilla.com/D207158
Some micro-optimization in the mozbuild/shellutil.py quoting function,
and use it at a finer grain to avoid some extra list creation.
This saves time in the configure step.
Differential Revision: https://phabricator.services.mozilla.com/D205861
We replace the use of FindFirstFileW by FindFirstFileExW with the most
optimized combination of flags on the code path reached from
nsIDirectoryEnumerator on Windows. This should make enumeration faster,
in particular when enumerating directories with thousands of files like
it can happen in QuotaManager code.
Differential Revision: https://phabricator.services.mozilla.com/D207151
- The inconsistencies were caused by the fact that the tabs weren't properly selected when trying to create a collection.
- To overcome this problem, I've added a retry loop to the the tab selection function and also added an extra assertion (**verifyTabsMultiSelectionCounter**()) inside the **selectTab** function which will help us identify more easily if there is a tab selection problem
- This patch should fix the following intermittent UI tests affected by the above mentioned problem:
- saveTabToExistingCollectionFromMainMenuTest
- renameCollectionTest
- verifyExpandedCollectionItemsTest
- removeTabFromCollectionUsingSwipeRightActionTest
- removeTabFromCollectionUsingSwipeLeftActionTest
- saveTabToExistingCollectionUsingTheAddTabButtonTest
✅ saveTabToExistingCollectionFromMainMenuTest - successfully passed 150x on Firebase
✅ renameCollectionTest - successfully passed 150x on Firebase
✅ verifyExpandedCollectionItemsTest - successfully passed 150x on Firebase
Differential Revision: https://phabricator.services.mozilla.com/D207146
The benefit is that whether the DeviceInputTrack or reverse stream could be
affected by clock drift is more likely to be known, and so recreation of
mAudioProcessing will be less likely when, in a subsequent patch, the
AudioProcessing needs to be replaced when the possibility of drift changes.
mSkipProcessing and mRequestedInputChannelCount are now replaced by their
MediaEnginePrefs equivalents. The AudioInputProcessing now starts in
pass-through mode, consistent with its initial AEC, AGC, and noise suppression
settings.
Differential Revision: https://phabricator.services.mozilla.com/D206868
Note that we intentionally return "null" for null principals *and* file
scheme, which is - at the time of writing - different than Chromium.
Differential Revision: https://phabricator.services.mozilla.com/D206989
Implements `PlacesBackupResource.backup` to store a copy of `places.sqlite` and `favicons.sqlite` in the staging folder.
If users don't want history remembered or use permanent private browsing mode, we will backup bookmarks instead to a file called `bookmarks.json`. Automatic backup is not yet implemented, so to test changes locally, go to `chrome://browser/content/backup/debug.html` to view where we store the staging folder and to manually run the backup methods for all available backup resources.
Backup files for `PlacesBackupResource` should be under the `places` subfolder in the staging folder.
Differential Revision: https://phabricator.services.mozilla.com/D206532
Added new secret data storing the release keys. Modifying scripts to hook them up. Testing will be done thoroughly with try and release builds to confirm that pinning works on newer machines and falls back to the old mechanism on older machines.
Differential Revision: https://phabricator.services.mozilla.com/D205361
This is adding in the new Windows 11 only version of taskbar pinning. For the new pinning to work, we need to use limited access feature tokens. Those are going to be made private and aren't included with this change.
This change will compile, and will work if built against the correct limited access feature tokens, as specified in developer local machine config files, but for every other build, the new taskbar pinning won't work and will fall back to the old methods.
I will implement the try / release building machines using the secret limited access feature tokens in a follow-up diff.
Differential Revision: https://phabricator.services.mozilla.com/D205004
In #621429 a call to nsUnixSystemProxySettings::GetSystemWPADSetting()
was added, but that function is not available for builds with
MOZ_ENABLE_LIBPROXY. The funtion is trivial enough to just
copy it across - else there'd more boilerplate than actual code.
Differential Revision: https://phabricator.services.mozilla.com/D207197
mPacketizerInput won't need to be re-created for every channel count change if
no processing is done for some channel counts.
More significantly, this will facilitate delaying, in a subsequent patch,
AudioProcessing construction while its parameters may still be changing.
Differential Revision: https://phabricator.services.mozilla.com/D206867
This clarifies that config, pass-through, and input channel count are all set
in the same way.
It will also facilitate delayed AudioProcessing construction in a subsequent
patch.
Differential Revision: https://phabricator.services.mozilla.com/D206866
Webpagetest has modified the way that they accept keys, they used to accept it through a URL parameter, but now they require it to be in the header and with a different name
Differential Revision: https://phabricator.services.mozilla.com/D207159
Without PREF_CUBEB_FORCE_SAMPLE_RATE in gInitCallbackPrefs, content processes
did not see the pref value unless the pref was changed during the lifetime of
the content process.
Differential Revision: https://phabricator.services.mozilla.com/D206864
