Bug 229790 – Add a layout test to detect memory leaks when drawing images to a detached canvas

• Component: Canvas
• Status: RESOLVED FIXED

[Said Abou-Hallawa]

This is a follow-up for bug 229668.

[Said Abou-Hallawa]

Created attachment 437106 [details]
Patch

[Said Abou-Hallawa]

Created attachment 437112 [details]
Patch

[Said Abou-Hallawa]

Without r281839 the test in this patch fails with the following diff:

- PASS internals.remoteImagesCount() is 0
+ FAIL internals.remoteImagesCount() should be 0. Was 3.

[Darin Adler]

Comment on attachment 437112 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=437112&action=review

> Source/WebCore/ChangeLog:3
> +        Add a layout test to detect memory leaks when darwing images to a detached canvas

darwing

> Source/WebKit/ChangeLog:3
> +        Add a layout test to detect memory leaks when darwing images to a detached canvas

darwing

> Source/WebCore/page/Page.h:584
> +    WEBCORE_EXPORT unsigned remoteImagesCount() const;

Given this is for testing only, could we have internals go directly to the chrome client instead of putting a helper in Page?

[Simon Fraser (smfr)]

> > Source/WebCore/page/Page.h:584
> > +    WEBCORE_EXPORT unsigned remoteImagesCount() const;
> 
> Given this is for testing only, could we have internals go directly to the
> chrome client instead of putting a helper in Page?

Let's rename it to remoteImagesCountForTesting() also.

[Said Abou-Hallawa]

Created attachment 437209 [details]
Patch

[EWS]

Committed r281980 (241287@main): <https://commits.webkit.org/241287@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 437209 [details].

[Radar WebKit Bug Importer]

<rdar://problem/82713895>

[Truitt Savell]

It looks like the test added here is failing fast/canvas/canvas-drawImage-detached-leak.html

History:
https://results.webkit.org/?suite=layout-tests&test=fast%2Fcanvas%2Fcanvas-drawImage-detached-leak.html


Diff:
--- /Volumes/Data/worker/bigsur-debug-tests-wk2/build/layout-test-results/fast/canvas/canvas-drawImage-detached-leak-expected.txt
+++ /Volumes/Data/worker/bigsur-debug-tests-wk2/build/layout-test-results/fast/canvas/canvas-drawImage-detached-leak-actual.txt
@@ -3,7 +3,7 @@
 On success, you will see a series of "PASS" messages, followed by "TEST COMPLETE".
 
 
-PASS internals.remoteImagesCountForTesting() is 0
+FAIL internals.remoteImagesCountForTesting() should be 0. Was 3.
 PASS successfullyParsed is true
 
 TEST COMPLETE

[Darin Adler]

That’s exactly what we’d expect if the fix in r281839 did not work.