Most of them are straight-forward. Some fuzziness due to Skia around rounded
corners still. Some tests that already fail in other widgets as expected.
This is enough to green up reftests (but not WPT, yet) on Android, both with and
without WR enabled.
Nothing too concerning, all-in-all.
Differential Revision: https://phabricator.services.mozilla.com/D65667
--HG--
extra : moz-landing-system : lando
Right now we post updates and it "works" because we prevent the UpdateOverflow
call if we're during reflow.
If this happens during styling however this is not sound (and it is not sound
in general and has caused badness in the past, as noted by the other
workarounds).
Make it sound by preventing to observe ancestors, and do it everywhere, removing
various ad-hoc hacks that were spread around elsewhere.
This changes expectations of two tests:
* clip-path-recursion-002.svg: Now we consider the inner clip-path reference
invalid. This matches WebKit and Blink, and I don't see any spec text
explicitly asking for our old behavior, so I just changed the test.
* element-paint-recursion.html: Changes the expectations of elements
referencing themselves via -moz-element(). Now it is invalid, instead of
painting ourselves once inside ourselves, which was a bit wild on its own.
Differential Revision: https://phabricator.services.mozilla.com/D53890
--HG--
extra : moz-landing-system : lando
These changes are needed for consistently green runs with the new emulator with
"-gpu on".
Most changes are simple removal of fuzzy-if(geckoview) but I also needed to add
at least one new fuzzy-if.
In this configuration we can run reftests in just 2 chunks (20 minutes each on
opt/30 minutes on debug).
Differential Revision: https://phabricator.services.mozilla.com/D36258
--HG--
extra : moz-landing-system : lando
There are a number of failures, for which I've filed separate bugs.
And then a lot of fuzziness. I manually inspected the reftest analyzer
results on try pushes to distinguish failures vs fuzziness.
Depends on D34537
Differential Revision: https://phabricator.services.mozilla.com/D34538
--HG--
extra : moz-landing-system : lando
Most rounded rect clips in real content are (a) axis aligned and
(b) have uniform radii.
When these conditions are met, we can run a fast path for clip
mask generation that uses significantly fewer ALU shader ops.
This is not typically a bottleneck on desktop GPUs, but can have
a large performance impact on mobile GPUs (and perhaps low end
integrated GPUs).
The Mali shader analyzer reports the slow path for the rounded
rect clip shader to be 94 cycles per fragment, while the fast
path is 10 cycles.
Differential Revision: https://phabricator.services.mozilla.com/D23817
--HG--
extra : moz-landing-system : lando
Most rounded rect clips in real content are (a) axis aligned and
(b) have uniform radii.
When these conditions are met, we can run a fast path for clip
mask generation that uses significantly fewer ALU shader ops.
This is not typically a bottleneck on desktop GPUs, but can have
a large performance impact on mobile GPUs (and perhaps low end
integrated GPUs).
The Mali shader analyzer reports the slow path for the rounded
rect clip shader to be 94 cycles per fragment, while the fast
path is 10 cycles.
Differential Revision: https://phabricator.services.mozilla.com/D23817
--HG--
extra : moz-landing-system : lando
Apparently we sometimes fall back to skia canvas with WebRender if ANGLE
fails to initialize. This makes the azureSkia condition true, but the
test still passes because of WebRender's magical properties. This patch
updates the annotation to reflect reality, so we don't get unexpected
passes.
Differential Revision: https://phabricator.services.mozilla.com/D21012
--HG--
extra : moz-landing-system : lando
Apparently we sometimes fall back to skia canvas with WebRender if ANGLE
fails to initialize. This makes the azureSkia condition true, but the
test still passes because of WebRender's magical properties. This patch
updates the annotation to reflect reality, so we don't get unexpected
passes.
Differential Revision: https://phabricator.services.mozilla.com/D21012
--HG--
extra : moz-landing-system : lando
add fuzzy-if statements to allow reftests to run on new windows10 AMI image
Differential Revision: https://phabricator.services.mozilla.com/D12553
--HG--
extra : moz-landing-system : lando
This is enough to fix the devtools regression and matches what other browsers
do in the no-attribute case.
Also, I think this change over all makes sense: it doesn't make any sense to
display the broken image icon if there's no request, and we already assume in
EnsureIntrinsicSizeAndRatio() that we don't paint the icon for those (and make
the intrinsic size 0x0).
We still show the border, which matches other UAs (note that devtools
effectively masks the border away with mask-image).
This technically also can change behavior of <object> and <input>, but I think
it's better to be consistent, since EnsureIntrinsicSizeAndRatio also doesn't
special-case <img> either.
Differential Revision: https://phabricator.services.mozilla.com/D11659
--HG--
extra : moz-landing-system : lando
This patch was written entirely by the following script:
#!/bin/bash
if [ ! -d "./.hg" ]
then
echo "Not in a source tree." 1>&2
exit 1
fi
find . -regex '.*\(ref\|crash\)test.*\.list' | while read FILENAME
do
echo "Processing ${FILENAME}."
# The following has four substitutions:
# * The first one replaces the *first* argument to fuzzy() when it doesn't
# have a - in it, by replacing it with an explicit 0-N range.
# * The second one does the same for the *second* argument to fuzzy().
# * The third does the same for the *second* argument to fuzzy-if().
# * The fourth does the same for the *third* argument to fuzzy-if().
#
# Note that this is using perl rather than sed because perl doesn't
# support non-greedy matching, which is needed for the first argument to
# fuzzy-if.
perl -pi -e 's/(fuzzy\()([^ ,()-]*)(,[^ ,()]*\))/${1}0-${2}${3}/g;s/(fuzzy\([^ ,()]*,)([^ ,()-]*)(\))/${1}0-${2}${3}/g;s/(fuzzy-if\([^ ]*?,)([^ ,()-]*)(,[^ ,()]*\))/${1}0-${2}${3}/g;s/(fuzzy-if\([^ ]*?,[^ ,()]*,)([^ ,()-]*)(\))/${1}0-${2}${3}/g' "${FILENAME}"
done
Differential Revision: https://phabricator.services.mozilla.com/D2974
--HG--
extra : moz-landing-system : lando
This covers all the reftests that have lower fuzz (or zero fuzz) and
were producing an UNEXPECTED-PASS result with webrender on windows. In
many cases I just adjusted the lower bound of the existing webrender
fuzz. In other cases existing fails-if conditions had to be tweaked to
exclude webrender.
MozReview-Commit-ID: 49LvS0vuYWR
--HG--
extra : rebase_source : d194e24affb87fe4560a127ff4016f9c38f414fd
Many of these could probably be fuzzed but in the interests of getting
the reftest suite turned on sooner I'm doing a blanket fails-if. This
covers all the reftests where there is more fuzz with webrender on
windows than any of existing annotations account for. In some cases the
fuzz is only a few pixels more than the equivalent Linux fuzz already
annotated, but I'll clean that up in a future bug.
MozReview-Commit-ID: IaKarbnL46d
--HG--
extra : rebase_source : 71889340305b0b12fa8eace722e42bb3faf14419
In these cases there is no human-visible difference in the WR renderings
of the test and reference images, so we can replace the failing
annotations with fuzzy ones instead.
MozReview-Commit-ID: Ioiem4ZwwI2
--HG--
extra : rebase_source : d2204d9674b4aba6eb78bc0d0007a40121564ac5
This patch:
- adds fails-if annotations for all the reftests that were consistently failing
with layers-free turned on.
- removes fails-if or reduces the range on fuzzy-if annotations for all
the reftests that were producing UNEXPECTED-PASS results with
layers-free turned on.
- adds skip-if, random-if, or fuzzy-if annotations to the reftests that
were intermittently failing due to timeout, obvious incorrectness, or
slight pixel differences, respectively.
MozReview-Commit-ID: A0Aknn6rnjj
--HG--
extra : rebase_source : 420d9cf43f23a5d654fa36eec69138937d13c173
Skip tests that are expected to fail in both Stylo and Gecko modes. They would unexpectedly "pass" in styloVsGecko mode when comparing the two failures, which is not a useful result.
MozReview-Commit-ID: 3mOpjU225Q1
--HG--
extra : rebase_source : 22bb5d4e3c5138ef832995eaf5716824f4707ffe
extra : source : d40fb20c9a49d0797c0eeae613a04912b12a28f7
Matt Woodrow
Cosmin Sabou
Cameron McCormack
Csoregi Natalia
Emilio Cobos Álvarez
Connor Brewster
Brian Grinstead
Brendan Dahl
Lee Salzman
Nicolas Silva
Gurzau Raul
Kartikaya Gupta
Coroiu Cristina
Geoff Brown
Joel Maher
Glenn Watson
Bogdan Tara
Ryan VanderMeulen
L. David Baron
Andreea Pavel
Jonathan Kew
Ethan Lin
Sebastian Hengst
Chris Peterson