Currently Gecko treats 'auto' component value as zero for animation.
MozReview-Commit-ID: JBvTFzDw7Xy
--HG--
extra : rebase_source : 618e756bbbb66759eea50a8004740e737f8a94e1
This gets us coverage for anonymous nodes, which are currently missed.
MozReview-Commit-ID: KBOSXloteu6
--HG--
extra : rebase_source : 0f71c027f8bbbb0abd4d9e812b5d7369109e7ebc
Currently Gecko treats 'auto' component value as zero for animation.
<!-- Please describe your changes on the following line: -->
https://bugzilla.mozilla.org/show_bug.cgi?id=1387951
---
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
Source-Repo: https://github.com/servo/servo
Source-Revision: e0b834033d857b08985fc84e676fac636a8495dc
--HG--
extra : subtree_source : https%3A//hg.mozilla.org/projects/converted-servo-linear
extra : subtree_revision : c4ba04702afbd72d77f23c1b8163d26998daefb7
Iterating a LiveRange's uses in BacktrackingAllocator::computeSpillWeight() to
access UsePosition.uses_ isn't cache friendly, because it is a linked list and
each node is allocated separately. Changing the data strcture to a vector is
worse because of the overhead from frequent insertion/removing. However,
UsePostion.uses_ can't be changed after initialization, so we can pre-calculate
the spill weight when it is added to or remove from a LiveRange.
MozReview-Commit-ID: BJEvI7KBVAJ
--HG--
extra : rebase_source : 04a14fae5a455d0e70f995e80887eb09b184dc2c
Copy construction for va_list is implementation dependent, and most
likely doesn't actually work properly on some platforms. While the
wrapping in EitherParser might currently work on tier-1 platforms,
it also triggers a warning on newer GCC versions that could very well
indicate that they might not work entirely properly. Mixing va_list,
copy construction and templates variadics surely sounds like something
very fragile.
While I understand that the pending changes from bug 1351107 might be
relying on the current state of the code, I'd rather that bug be
reworked so as to please newer versions of GCC than having to figure out
what parts of that bug are still relevant after bug 1351105 and blocking
bug 1389435 further.
--HG--
extra : rebase_source : b9135990c62faf7587e646c99f3bc6ece603cb38
This requires adding a new constructor for ns[C]String that can be used to
create an IsVoid string.
--HG--
extra : rebase_source : 8cb078bd0a41e63af0d9d144b9eef369875a05e0
Web contetnt processes only need access to a small amount of schema data, but
we currently send them the approximately 600K of full schema data that is
mostly useless to them.
This patch limits the schema data sent to web content processes to what they
actually need, and sends the rest only to extension content processes.
MozReview-Commit-ID: 6G0LThNTOu1
--HG--
extra : rebase_source : 36672ad6323e6466bba3e463fa4f0a16e3fd9090
This gives JS callers access to the remote type of remote message managers.
There's currently no way for extensions to access this unless they have a
<browser> element to check the remoteType attribute of.
MozReview-Commit-ID: A8Y3ZSG3rt8
--HG--
extra : rebase_source : e024922522da9a30265f05e9a8dbf7529dfe1d81
JS code is notified when a new ContentParent is created via normal
"ipc:content-created" notifications, but can't do anything with it, since
nsIContentParent is not scriptable. This allows JS callers to retrieve the
parent process message manager, which is the normal way they interact with
content children.
MozReview-Commit-ID: 7lcZ4XkJ6uR
--HG--
extra : rebase_source : f891c0e29863fc42fc2351a791ca3f1f7e2824b9
The webdriver spec has been updated a while ago and renamed the proxyType
for not using a proxy from "noProxy" to "direct".
MozReview-Commit-ID: FjDIK8XCba8
--HG--
extra : rebase_source : 423149d3f82cf1b7e28d29a775359ac66e7a6f90
When the tooltool manifest contains e.g. clang.tar.bz2 and it's also
provided by some toolchain job artifact, we still download both. In
fact, depending how things go, the tooltool one could take precedence.
In practice, this hasn't caused problems so far because we've removed
tooltool manifest items early on, and when we didn't, it was still the
same version as provided by toolchain jobs.
It is still necessary to fix for the future, though.
--HG--
extra : rebase_source : 1d187333ee6149f72e3f9ed91eb1a6a78ad9197f
To allow geckodriver to create file objects in child processes, this
preference has to be set to true. As such it needs to be recommended
in Marionette and set as default preference for geckodriver.
MozReview-Commit-ID: 8bdbi5jH6Vc
--HG--
extra : rebase_source : cc8963d2b15357b6d1607854e978aa764fcbd42f
<!-- Please describe your changes on the following line: -->
Required to land https://github.com/emilio/rust-offscreen-rendering-context/pull/105
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `__` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [ ] These changes fix #__ (github issue number if applicable).
<!-- Either: -->
- [x] There are tests for these changes OR
- [ ] These changes do not require tests because _____
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Source-Repo: https://github.com/servo/servo
Source-Revision: e9cbbc58cc9655a15bc603effabbd4a4ff62b454
--HG--
extra : subtree_source : https%3A//hg.mozilla.org/projects/converted-servo-linear
extra : subtree_revision : f76bcef3944751922ccdd1e20a462105d5806a7b
These getters are checked very rarely, and not at all in most sessions. They
don't justify the overhead of adding lazy getters at startup.
MozReview-Commit-ID: 9XVlLapNJCE
--HG--
extra : rebase_source : edff8e878528952aeec851203edaa4d41e37e24d
The webdriver spec has been updated a while ago and renamed the proxyType
for not using a proxy from "noProxy" to "direct".
MozReview-Commit-ID: FjDIK8XCba8
--HG--
extra : rebase_source : ab41d737c9ba06c8883bfd36839807e6c3316c92
CompareCacheHashEntry::mCrit[] is the only case where the nsXPIDLString-ness
was important. The patch adds an explicit SetIsVoid() call to that class's
constructor and changes some null checks to IsVoid() checks.
--HG--
extra : rebase_source : e68befcde4dd098bac2a550bc666eaf3bf1298d7
Kris Maglione
Bill McCloskey
Phil Ringnalda
Hiroyuki Ikezoe
Bevis Tseng
Nicholas Nethercote
Wes Kocher
Botond Ballo
Ting-Yu Chou
Mike Hommey
Jonathan Kingston
Henrik Skupin
Edouard Oger
Sebastian Hengst
Aki Sasaki
Kartikaya Gupta
Imanol Fernandez
Mason Chang
Ryan Hunt
Gijs Kruitbosch
Aryeh Gregor