Repro'd the earlier failure (sorry), and adjusted the screensize to 1024x768 in the test fixture. I also added the linux-wayland-weston-rel-tests bot, which now passes. Would like to try this again, thanks for considering.
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
Gerrit GetPureRevert API does not mark this CL as a pure revert. Learn more: go/rubber-stamper-user-guide.
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
| Commit-Queue | +2 |
Thanks a lot. I'm submitting this quickly so we can see soon whether the small adjustment is sufficient.
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
| Code-Review | +1 |
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
Thanks for the investigation. The reland doesn't need gardener's review, so I'm removing myself from the reviewer list.
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
Reland "[soft navs] Improve CLS test coverage."
This reverts commit a49d7dde42459da22036a6da1c473a0fa4c58884.
Reason for revert: Passes w/ adjusted screen size (see patchset 2).
Original change's description:
> Revert "[soft navs] Improve CLS test coverage."
>
> This reverts commit b12db17c1ee1f4644cdf3ef7dd90769ea20e957c.
>
> Reason for revert: caused timeout issue in linux-wayland-weston-rel-tests. https://ci.chromium.org/ui/p/chromium/builders/ci/linux-wayland-weston-rel-tests/14104/overview
>
> Original change's description:
> > [soft navs] Improve CLS test coverage.
> >
> > * Add coverage for softnav CLS after prerender and bfcache restores.
> > * Migrate CLS coverage for regular page loads
> > (SoftNavigationTest.LayoutShift) to use
> > soft_navigation_basics.html, instead of soft_navigation.html, as
> > it's drastically simpler.
> > * Due to the migration, SoftNavigationTest::GetCLSFromList is no
> > longer needed. Note that the replacement does not perform the
> > session window / gap algorithm, instead, it assumes that some
> > adjacent layout shifts are going to be expressed as their sum
> > in UKM. I think this is good enough, as the window / gap stuff
> > is tested elsewhere.
> >
> > For now, I'm doing exact comparisons (EXPECT_EQ) between the ukm CLS
> > values and the values coming from the renderer. I'm hoping that we can
> > get away with this and not see flakes (haven't seen any thus far), even
> > though the values from the renderer may have added noise and get
> > converted to integral values by rounding. But the alternative with
> > converting to double and EXPECT_NEAR like it's done in
> > layout_instability_browsertest.cc etc. seems less appealing because the
> > typical 1.0 allowed error may end up missing salient differences.
> >
> > Bug: 496610192
> > Change-Id: I2e77ce9ff8085f3677bbe8636bbc621609cca80d
> > Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/7842929
> > Reviewed-by: Scott Haseley <shas...@chromium.org>
> > Commit-Queue: Johannes Henkel <joha...@chromium.org>
> > Reviewed-by: Annie Sullivan <sull...@chromium.org>
> > Cr-Commit-Position: refs/heads/main@{#1632379}
>
> Bug: 496610192
> No-Presubmit: true
> No-Tree-Checks: true
> No-Try: true
> Change-Id: I44ff2a159eaad8a9fbcdc1e36c7b0e18dbeb672a
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/7857620
> Auto-Submit: Xinghui Lu <xing...@chromium.org>
> Owners-Override: Xinghui Lu <xing...@chromium.org>
> Commit-Queue: rubber-...@appspot.gserviceaccount.com <rubber-...@appspot.gserviceaccount.com>
> Bot-Commit: rubber-...@appspot.gserviceaccount.com <rubber-...@appspot.gserviceaccount.com>
> Cr-Commit-Position: refs/heads/main@{#1632462}
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |