The mozharness scripts have a lot of special case arguments for one off
configurations, stuff like --e10s, --enable-webrender and --gpu-required.
Many of these command line args ultimately only end up setting an extra pref in
the test harnesses. Instead, let's just give mozharness the ability to set
prefs directly via --setpref. This way we can pass them through from taskgraph
without needing to add extra configuration to mozharness when making changes
like this.
Differential Revision: https://phabricator.services.mozilla.com/D7191
--HG--
extra : moz-landing-system : lando
This emits a warning when the webdriver library reverts to using the
legacy Selenium-styled capabilities negotiation for new sessions.
It provides a helpful link which will explain to users how to update.
Differential Revision: https://phabricator.services.mozilla.com/D7305
--HG--
extra : moz-landing-system : lando
This test runs multiple testcases, so sometimes 10 seconds are not enough for
all of them to finish. This commit bumps the timeout to 60 seconds and reenabled
the test on Linux.
Differential Revision: https://phabricator.services.mozilla.com/D7291
--HG--
extra : source : 9a685b824b8862df011d96b6a246f7ff83fe5e67
extra : amend_source : ebcbba770e232cef4ae20e15fbaea48dedd3e02a
This test runs multiple testcases, so sometimes 10 seconds are not enough for
all of them to finish. This commit bumps the timeout to 60 seconds and reenabled
the test on Linux.
Differential Revision: https://phabricator.services.mozilla.com/D7291
--HG--
extra : moz-landing-system : lando
We need to sign parts of the contents of the archives, so the mar's that we
ship get built as part of the repackage task. Thus, there is no reason to also
create and upload as part of the build, just to throw them away.
Differential Revision: https://phabricator.services.mozilla.com/D6213
--HG--
extra : moz-landing-system : lando
This test tried to clean up after itself, but didn't do a very good
job at it. I didn't analyze it fully, but I suspect that the ordering
of the microtasks in the test depended on a race condition that caused
it to fail from time to time. This commit moves cleanup code to the
test harness, which does a better job at ensuring that things happen
in order.
Differential Revision: https://phabricator.services.mozilla.com/D7278
--HG--
extra : moz-landing-system : lando
tresize-test.html attempts to load
"resource://talos-powers/TalosPowersContent.js", but this might not be available
yet if the tresize add-on hasn't loaded. This patch changes the location of this
resource to a relative path, which should always be available. Additionally,
this patch awaits on TalosPowers.loadPromise before using talos APIs.
Differential Revision: https://phabricator.services.mozilla.com/D7312
--HG--
extra : moz-landing-system : lando
Add support for verify and wasm test info when running web platform tests on
android
Differential Revision: https://phabricator.services.mozilla.com/D7270
--HG--
extra : moz-landing-system : lando
Multiprocessing on Linux and OSX implictly inherits the stdio handles
into the subprocess. But on Windows that doesn't happen because it
doesn't have fork() to abuse in the same way. So instead we need to
ensure that we explicitly set up handling of stdio immediately after
starting the executor subprocess to avoid gettin non-structured output.
Differential Revision: https://phabricator.services.mozilla.com/D7186
--HG--
extra : moz-landing-system : lando
Before we were using a different logger for the manifest download and the actual test run.
This caused timestamps to get reset in a confusing way. Now create the logger early and
share it for all the subseteps.
Depends on D7171
Differential Revision: https://phabricator.services.mozilla.com/D7172
--HG--
extra : moz-landing-system : lando
As far as I know this was unused and hasn't been well maintained since it's
not that useful.
Differential Revision: https://phabricator.services.mozilla.com/D7171
--HG--
extra : moz-landing-system : lando
The patch c1df1c2e46f6 contained a faulty rebase where the coercion
of logging::max_level() changed from the Pref type to a string.
The string representation of geckodriver::logging::Level is given
in upper-case, e.g. "INFO", and the Pref representation is given as
"Info" to be compatible with managing the log level from Log.jsm in Gecko.
This inadvertently caused
https://github.com/web-platform-tests/wpt/issues/12166 to reappear
in almost the same way: before the problem was that Marionette’s
frame script always included all log level entries. This was fixed
with https://bugzilla.mozilla.org/show_bug.cgi?1482829, but then
https://bugzilla.mozilla.org/show_bug.cgi?id=1396821 broke it so
that log entries also from chrome space appeared.
mozprofile gives us an error and we can forward it to the user so
we know what the underlying reason is for not being able to write
the Firefox profile.
Thunderbird releases come from the comm-esr.* repositories rather than the
comm-release repository. This adds a special case for Thunderbird and sets
the branch URL correctly.
This initial patch is based on the branch_prefix (comm or mozilla).
Differential Revision: https://phabricator.services.mozilla.com/D6509
--HG--
extra : moz-landing-system : lando
Automatic update from web-platform-testsAdd headless mode for chrome_webdriver (#13201)
0200c63a74a540dfb870881d5f476280c367d1c1 failed to add this to
chrome_webdriver, so add this now.
--
wpt-commits: 1e7408fcd394b66234c38c24ba044ec6f00e4b70
wpt-pr: 13201
Automatic update from web-platform-tests[css-properties-values-api] Substitution behavior.
Currently, registered properties substitute into other values exactly
like specified (like unregistered properties). This means that, for
a <length>-registered property with a specified value "10em" (for
instance), when that property is substituted via var()-reference,
the tokens "10em" are inserted. This is not correct, and produces the
wrong result with e.g. inherited values.
This CL changes that, by implementing "absolutization" of registered
custom properties: a process which calculates the computed value of
the property, and then produces a token stream usable for substitution
which is equivalent to the computed value.
* Currently we resolve var()-references on all custom properties
before applying high-priority properties. This is no longer
possible, because the true value of a custom property (also
unregistered) can not be known until the font has been updated.
Consider: --reg-len: 1em; --unreg:var(--reg-len). Here, the
computed value of --unreg should be "16px" (assuming a font-size
of 16px), and not "1em". If we destructively resolve --unreg
before the font size is known, we end up with the wrong tokens.
Hence the resolution process has been moved to after the font
has been updated.
* Custom properties must also be usable from high-priority properties.
Since they are no longer resolved beforehand, they are resolved
"on the fly", non-destructively. "Non-destructively" means that
resolved token streams are not stored on ComputedStyle, such that
any var()-references are kept for the _real_ resolution pass after
the font has been updated.
This works, because the high-priority properties do not require
proper "absolute substitution" to produce the correct value.
Note that font-size is special, in that font-relative units may
not be used if they arrive via a registered custom property.
* There is a special resolving pass for registered custom properties
(ComputeRegisteredVariables). This pass produces non-absolute CSSValues
for calculation of animation.
* Parsing of registered properties now happens entirely in
CSSVariableResolver, and no longer in variable.cc. Having this in
multiple places is just confusing.
R=futhark@chromium.org
Bug: 641877
Change-Id: Ic705d0808ffcea0ae5db02fb20870767175bb706
Reviewed-on: https://chromium-review.googlesource.com/1240274
Reviewed-by: Rune Lillesveen <futhark@chromium.org>
Commit-Queue: Anders Ruud <andruud@chromium.org>
Cr-Commit-Position: refs/heads/master@{#593902}
--
wpt-commits: c801e269e3e22d11e1bb4530ff15e87db877b1d2
wpt-pr: 13193
Automatic update from web-platform-testsFix firefox find_binary_path to accomodate channels
Firefox find_binary_path returned None since browser-channels were introduced
due to path differences. This fixes the same, along with a minor change to the
Nightly path on Mac as it contains a space.
The default for |wpt run firefox| without a channel specified is set to nightly,
along with a log message when this happens.
--
wpt-commits: f3a1f9324097311cafcd7893066e3e40fe172fd0
wpt-pr: 13185
Automatic update from web-platform-testsAdd interface tests for TransitionEvent
These are modelled after the similar tests that exist for AnimationEvent
in css/css-animations, except written using more modern WPT style.
Bug: 695504
Change-Id: I63573cf48aba6821f5e22486ef18217d0f2828ee
Reviewed-on: https://chromium-review.googlesource.com/1239457
Reviewed-by: Stephen McGruer <smcgruer@chromium.org>
Reviewed-by: Xida Chen <xidachen@chromium.org>
Commit-Queue: Stephen McGruer <smcgruer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#593620}
--
wpt-commits: b040cb2a07264d35b9f47cd9deb2fb2889fca6d4
wpt-pr: 13168
Automatic update from web-platform-tests[PE] Fix fragment clip and paint offset under multicol vertical-rl scrolled
The previous code in VisualOffsetFromPaintOffsetRoot() converting the result
into scrolling contents space used ScrolledContentOffset which mismatched
the new ScrollTranslation which also includes scroll origin. Now use
ScrollTranslation instead.
Also removed the wrong remedy about scroll origin in one of the callers.
Bug: 887423
Cq-Include-Trybots: luci.chromium.try:linux_layout_tests_slimming_paint_v2;master.tryserver.blink:linux_trusty_blink_rel
Change-Id: I9a72ed6ddc066c5b4c9970037ed4c4287c5a2f0f
Reviewed-on: https://chromium-review.googlesource.com/1239170
Commit-Queue: Xianzhu Wang <wangxianzhu@chromium.org>
Reviewed-by: Morten Stenshorne <mstensho@chromium.org>
Reviewed-by: Philip Rogers <pdr@chromium.org>
Cr-Commit-Position: refs/heads/master@{#593572}
--
wpt-commits: 68d63d516e6bc4b83934d71616b2aa82d8efe2af
wpt-pr: 13157
Automatic update from web-platform-tests[css-grid] Performance optimizations in the Baseline alignment code
Since we integrated the baseline alignment logic in the grid tracks
sizing algorithm, its impact on performance has grown considerably.
The analysis of the new logic added and its overhead, due to different
operations, shows that evaluating the item's participation in the
baseline alignment context is the most expensive one. It's specially
demanding the evaluation of the grid item's alignment properties.
Considering that currently we are doing this for every grid item, this
CL propose to reuse the loop we already have to clear the grid item's
override size to cache the items with a baseline value in their
alignment CSS properties.
Thanks to this cache we can determine the item's participation in a
baseline alignment context in the different phases of the track sizing
algorithm, with almost no cost (compared to the current logic). It may
be possible to share the cache with the algorithm run for computing the
grid's intrinsic size; however, if the intrinsic size logic is run
before the grid's layout, we'll end up duplicating the cache.
Additionally, this cache is also used in the alignment phase of the
grid layout logic; this change helps to avoid the various issues we
have been suffering related to the different evaluations of the item's
participation in baseline during the different phases of the grid
layout algorithm.
BUG = 873452
Change-Id: Ida27be11ae0f5c455e6077367a277981ab35cec1
Reviewed-on: https://chromium-review.googlesource.com/1179897
Commit-Queue: Javier Fernandez <jfernandez@igalia.com>
Reviewed-by: Manuel Rego <rego@igalia.com>
Cr-Commit-Position: refs/heads/master@{#593552}
--
wpt-commits: f6de5f7d5d50b1ad30a1e17922e183868fd3e935
wpt-pr: 12816
Automatic update from web-platform-testsAdd a headless mode command line option (#13076)
For browsers which support headless mode starting wptrunner with
--headless will run the tests in headless mode; otherwise it will be
ignored.
--no-headless is avaiable to explicitly disable the mode when
frontends change the defaults.
For firefox switch the default for wpt run to be headless; for Chrome
changing the default broke some smoketests, so don't switch until that
is resolved. In all cases continue to run in a normal windowed mode in CI.
--
wpt-commits: 0200c63a74a540dfb870881d5f476280c367d1c1
wpt-pr: 13076
Automatic update from web-platform-testsVisual change on background-clip changes.
The background-clip property affects background-color visually, so we
need to say a layer is visually different even without an image.
We had a wpt test for this which had a stray line which made the test
always pass.
Bug: 887350
Change-Id: Icc585f9e631485970b6509e015eff1e2007e4f0b
Reviewed-on: https://chromium-review.googlesource.com/1238448
Reviewed-by: Anders Ruud <andruud@chromium.org>
Reviewed-by: Fredrik Söderquist <fs@opera.com>
Commit-Queue: Rune Lillesveen <futhark@chromium.org>
Cr-Commit-Position: refs/heads/master@{#593523}
--
wpt-commits: 00c1fd9ea4bc1e03f966919f1e209ab8cd81f57e
wpt-pr: 13187
Automatic update from web-platform-testsFix some tests in test_request.py for python 3. (#11757)
--
wpt-commits: 61e09bec3d21b4e9137dadafc20c04cb9be7ce3e
wpt-pr: 11757
Automatic update from web-platform-testsUpdate mozrunner from 7.0.0 to 7.0.2 (#13167)
--
wpt-commits: 91491deb7dcfe4a0d0ece61f086345abad4d46a4
wpt-pr: 13167