Bug 214388 – [macOS WK1 EWS] compositing/repaint/iframes tests are flaky failures

• Status: NEW
• Component: Compositing

[Ryan Haddad]

compositing/repaint/iframes/composited-iframe-with-fixed-background-doc-repaint.html is a flaky failure on macOS WK1 EWS bots

--- /Volumes/Data/worker/macOS-Mojave-Release-WK1-Tests-EWS/build/layout-test-results/compositing/repaint/iframes/composited-iframe-with-fixed-background-doc-repaint-expected.txt
+++ /Volumes/Data/worker/macOS-Mojave-Release-WK1-Tests-EWS/build/layout-test-results/compositing/repaint/iframes/composited-iframe-with-fixed-background-doc-repaint-actual.txt
@@ -14,6 +14,11 @@
           (drawsContent 1)
           (repaint rects
             (rect 16.00 16.00 400.00 300.00)
+            (rect 16.00 16.00 400.00 300.00)
+            (rect 24.00 24.00 100.00 100.00)
+            (rect 24.00 24.00 384.00 292.00)
+            (rect 16.00 16.00 400.00 300.00)
+            (rect 16.00 16.00 400.00 300.00)
             (rect 24.00 16.00 100.00 8.00)
           )
         )

Examples
https://ews-build.webkit.org/#/builders/30/builds/13889
https://ews-build.webkit.org/#/builders/30/builds/13884
https://ews-build.webkit.org/#/builders/30/builds/13883

Test history (no failures on post-commit macOS bots as of this filing:
https://results.webkit.org/?suite=layout-tests&test=compositing%2Frepaint%2Fiframes%2Fcomposited-iframe-with-fixed-background-doc-repaint.html

[Radar WebKit Bug Importer]

<rdar://problem/65636360>

[Ryan Haddad]

Marked test as flaky in r264438

[Ryan Haddad]

Since this isn't failing on post-commit bots, we need to try to run tests locally with `--skip-failing-tests` or a test list to get reproduction steps.

[Karl Rackler]

I am able to reproduce this for r264130, but not with r264129

REPRODUCTION STEPS
run-webkit-tests --iterations 2000 --exit-after-n-failures 3 --exit-after-n-crashes-or-timeouts 1 --debug-rwt-logging --no-retry --force --no-build -f -1 compositing/repaint/iframes/composited-iframe-with-fixed-background-doc-repaint.html
Diff:
--- /Volumes/Data/tmp/MacRelease/layout-test-results/compositing/repaint/iframes/composited-iframe-with-fixed-background-doc-repaint-expected.txt
+++ /Volumes/Data/tmp/MacRelease/layout-test-results/compositing/repaint/iframes/composited-iframe-with-fixed-background-doc-repaint-actual.txt
@@ -14,6 +14,11 @@
           (drawsContent 1)
           (repaint rects
             (rect 16.00 16.00 400.00 300.00)
+            (rect 16.00 16.00 400.00 300.00)
+            (rect 24.00 24.00 100.00 100.00)
+            (rect 24.00 24.00 384.00 292.00)
+            (rect 16.00 16.00 400.00 300.00)
+            (rect 16.00 16.00 400.00 300.00)
             (rect 24.00 16.00 100.00 8.00)
           )
         )

[Simon Fraser (smfr)]

(In reply to Karl Rackler from comment #4)
> I am able to reproduce this for r264130, but not with r264129

That doesn't really make sense. r264130 is a test-only change which did not affect this test or any of its results.

[Ryan Haddad]

Disabling this test caused compositing/repaint/iframes/compositing-iframe-scroll-repaint.html to start failing, and disabling that test caused compositing/repaint/iframes/compositing-iframe-with-fixed-background-doc-repaint.html to start failing, so we really need to figure out what test is causing all this.

[Ryan Haddad]

Marked the compositing/repaint/iframes directory as failing for Mojave WK1 in r264840 in an effort make EWS happy.