It's not worth dying on this hill. Both Blink and WebKit pass the tests.
(Well, WebKit actually fails one of the latest ones I wrote, cssRules and rules
are not the same JS object, WebKit returns a new rule list. I'll file)
Spec PR in https://github.com/w3c/csswg-drafts/pull/3900.
Differential Revision: https://phabricator.services.mozilla.com/D30348
--HG--
extra : moz-landing-system : lando
The "Make default browser" setting is now just a button, part of the "General" settings.
We'll remove the layout for 2-pane settings and the now unused String.
Depends on D30468
Differential Revision: https://phabricator.services.mozilla.com/D30469
--HG--
extra : moz-landing-system : lando
Update this junit test with this bug's test case. Actually, autofill_userpass
doesn't work with the latest GV, so I should like to update this to fix focus
timing.
Also, this does't run on our test infra because this requires API 26.
Differential Revision: https://phabricator.services.mozilla.com/D30178
--HG--
extra : moz-landing-system : lando
Actually, sample url has `<fieldset>` element, so LoginManagerContent cannot detects username field.
`LoginManagerContent.getUserNameAndPasswordFields` detects username field and password field when passing `<input>` element. So we should pass password field if possible.
Differential Revision: https://phabricator.services.mozilla.com/D30179
--HG--
extra : moz-landing-system : lando
Flip pref in test_basic_form_autocomplete.html to allow eval() temporarily.
Differential Revision: https://phabricator.services.mozilla.com/D30016
--HG--
extra : moz-landing-system : lando
This patch fixes some build issues on Android and also enables some gtests.
Differential Revision: https://phabricator.services.mozilla.com/D30386
--HG--
extra : moz-landing-system : lando
`nsIDocumentStateListener` is a scriptable interface and each method may run
any script. So, we should mark them as `can_run_script`. Then, we need to
mark a lot of editing methods because we need to mark
`EditorBase::EndTransactionInternal()` and `EditorBase::DoTransactionInternal()`
as `MOZ_CAN_RUN_SCRIPT`.
Differential Revision: https://phabricator.services.mozilla.com/D30360
--HG--
extra : moz-landing-system : lando
When using `CheckAsync` IPC, we don't hold `mozSpellChecker`. It causes that
`mozSpellChecker` may be destroyed during IPC call.
The destructor of `mozSpellChecker` destroys actor of spellchecker IPC via
`Send__delete__`. Although IPC rejects pending promises of async IPC during
destroying actor, it cannot remove replay messages from parent process.
So route error occurs.
So we have to keep `mozSpellChecker` during async IPC.
And we cannot convert attached test case to crashtest or mochitest since this
depends on navigation and timing. So I don't add it.
Differential Revision: https://phabricator.services.mozilla.com/D30347
--HG--
extra : moz-landing-system : lando
As per bug 1353918 comment 13, all these properties are animatable on the Servo
side so we no longer need this check.
Differential Revision: https://phabricator.services.mozilla.com/D30458
--HG--
extra : moz-landing-system : lando
In test0(), we use the callback of `requestAnimationFrame` to know in which
eventloop (i.e. `Tick()`) we are. However, we may not trigger the callback
of `requestAnimationFrame` if we are not visible. This is an optimization
in Bug 1145439.
Detail:
We use `Document::ShouldThrottleFrameRequests()` to check if we should throttle
the frame requests in the current `Tick()`. This function returns true if we
didn't get painted during the last paint, so we are not visible, so throttle
the frame requests. Note that because we have to paint this document at least
once to unthrottle it, we will drop one `requestAnimationFrame` frame when a
document that previously wasn't visible scrolls into view.
Therefore, we should make sure we got the first paint before running test0().
Using onload is not perfect, but we don't have other better choose for now.
Differential Revision: https://phabricator.services.mozilla.com/D29772
--HG--
extra : moz-landing-system : lando
Add multicol-width-004.html and multicol-width-005.html to test "width:
min-content" and "width: max-content" with column-span:all children.
There's no size containment in these tests.
Note it may be worth to reuse nsBlockFrame's mCachedMinISize and
mCachedPrefISize to cache intrinsic size for ColumnSetWrapperFrame, but
this can be done separately.
Differential Revision: https://phabricator.services.mozilla.com/D29616
--HG--
extra : moz-landing-system : lando
If column-span is enabled, nsColumnSetFrame becomes an inner frame under
ColumnSetWrapperFrame, and that outer frame is then where we should be
handling size containment.
This patch doesn't change the logic but make the logic easier to
simplify once we start eliminating column-span pref.
Differential Revision: https://phabricator.services.mozilla.com/D29615
--HG--
extra : moz-landing-system : lando
The functions move to ColumnUtils will be used by ColumnSetWrapperFrame
to implement GetPrefISize() and GetMinISize().
Also, I verify locally that non-unified build is still working by
s/UNIFIED_SOURCES/SOURCES/ in layout/generic/moz.build.
Differential Revision: https://phabricator.services.mozilla.com/D29614
--HG--
extra : moz-landing-system : lando
This ensures that the iframe is loaded by the time we query its visual viewport size.
Differential Revision: https://phabricator.services.mozilla.com/D30417
--HG--
extra : moz-landing-system : lando
This causes some scrollbar-related assertions to fail on desktop because
scrollbar behaviour with desktop zooming is not correct yet.
Differential Revision: https://phabricator.services.mozilla.com/D30415
--HG--
extra : moz-landing-system : lando
1. Add new tests with an extra argument of "content-box"/"border-box" in
observe(), and test contectBoxSize and borderBoxSize.
2. Add a new test for changing the writing mode. Changing writing mode makes
the element change the orientation, but the logical box size is still
the same, so we shouldn't fire the event.
Differential Revision: https://phabricator.services.mozilla.com/D28739
--HG--
extra : moz-landing-system : lando
I think it's better to make sure each test is independent from others, so we
create elements inside each test function. (Only change observe.html
because I touch only this file in this bug.)
Differential Revision: https://phabricator.services.mozilla.com/D29889
--HG--
extra : moz-landing-system : lando
In this patch, we support
1. content-box (default)
2. border-box
And let ResizeObserverEntry expose these box sizes.
Besides, we store the mLastReportedSize as the logical size.
Differential Revision: https://phabricator.services.mozilla.com/D28737
--HG--
extra : moz-landing-system : lando
GetNodeDepth() is a special version for ResizeObserver to get the depth
of node (across Shadow DOM). Based on the comment in D27615, it's better
to move it into ResizeObserver.cpp.
Differential Revision: https://phabricator.services.mozilla.com/D28736
--HG--
extra : moz-landing-system : lando
An assert occurs while calculating durations whenever we reach the end of the event list and we didn't encounter the TraceLogger_Stop events for events that were still active when the profiler stopped. The fix is to use the last event as the end event for any remaining events on the stack.
Differential Revision: https://phabricator.services.mozilla.com/D29926
--HG--
extra : moz-landing-system : lando
