Given that we would otherwise disable remote canvas when we disable the
GPU process, it is not meaningfully worse for users to just disable
remote canvas keeping the GPU process as the final step. It is still
preferable to losing the GPU process sandbox. This includes D2D canvas,
accelerated canvas and WebGPU, as well as WebGL on Android.
Original Revision: https://phabricator.services.mozilla.com/D267667
Differential Revision: https://phabricator.services.mozilla.com/D267862
Before this patch, an attempt to read from a program that closed its
stdout would result in a never-resolving read promise on Windows.
Before this patch, an attempt to write to a program that closed its
stdin would result in a never-resolving write promise on Windows.
This patch fixes both issues by fixing the typo that prevented the
implementation from detecting errors, and adds test coverage.
Original Revision: https://phabricator.services.mozilla.com/D261332
Differential Revision: https://phabricator.services.mozilla.com/D267841
This lets us encode updated expectations for these scenarios (based on more
web=compatible soon-to-be-proposed spec text[1]) right now in
naturalWidth-naturalHeight-width-height.tentative.html, while letting
naturalWidth-naturalHeight.html continue to encode the expectations of current
spec text for the time being.
(Note that some of the scenarios in the older test are already tested in the
newer test and hence don't need to be copied over.)
[1] https://github.com/whatwg/html/issues/11287
Original Revision: https://phabricator.services.mozilla.com/D251970
Differential Revision: https://phabricator.services.mozilla.com/D267121
I originally created this test as a placeholder to test some edge cases where
the naturalWidth and naturalHeight might be unavailable (hence "unavailable" in
the name), with the intent of ultimately merging it back into
naturalWidth-naturalHeight.html once
https://github.com/whatwg/html/issues/11287 was resolved.
However, I'm now planning for this test to just be the canonical test
for the img.{naturalWidth,naturalHeight,width,height} attributes, for a variety
of scenarios (not just ones where the natural width/height might be
unavailable). Then we can remove the naturalWidth-naturalHeight.html test
entirely (obsoleted by this test) when the spec change is complete.
This patch is the first part of this -- just renaming the test and updating
the title to reflect its new broader purpose.
Original Revision: https://phabricator.services.mozilla.com/D251969
Differential Revision: https://phabricator.services.mozilla.com/D267120
This is just cleanup and doesn't change the behavior of this test. The img
elements here are 'display:block', so:
* 'vertical-align' doesn't apply.
* 'width:max-content' is basically redundant, since block-level replaced
elements already size to their max-content size.
Hence, I'm removing these unnecessary styles.
(I re-ran the test in both Firefox and Chrome to make sure the results don't
change.)
Original Revision: https://phabricator.services.mozilla.com/D251934
Differential Revision: https://phabricator.services.mozilla.com/D267119
This patch doesn't fundamentally change the behavior or expectations of this
test.
This patch simply moves all of the img elements to a <template> (which is not
rendered), and then appends a clone of those elements to a container within
the body; and then those cloned img elements are what get tested.
This sets the stage for later patches to append additional modified copies of
these img elements, with small modifications (e.g. using 'srcset' rather than
'src').
Original Revision: https://phabricator.services.mozilla.com/D251659
Differential Revision: https://phabricator.services.mozilla.com/D267118
While reverting all the work related to cross platform escaping of curl (See Bug 1976589), i tried to cleanup the
escaping of new line characters, but introduced carriage returns (\r).
This patch reverts back to the original fix from Bug 1968414
Original Revision: https://phabricator.services.mozilla.com/D265397
Differential Revision: https://phabricator.services.mozilla.com/D267045
This patch reverts important parts of D243110 related to Bug 1950001
This goal is for the Copy As Curl(Windows) to work properly on just the windows CMD
Differential Revision: https://phabricator.services.mozilla.com/D255793
- This patch should revert the fixes from Bug 1960198, Bug 1949994 and Bug 1962301.
- The goal of this patch is Copy as curl (POSIX) should only work on Linux.
Differential Revision: https://phabricator.services.mozilla.com/D255784
