Bug 9115

Summary:

Content & layers with Opacity 0 are rendered (invisibly) when they don't need to be

Product:

WebKit

Reporter:

David Carson <dacarson>

Component:

Layout and Rendering

Assignee:

David Carson <dacarson>

Status:

RESOLVED FIXED

Severity:

Enhancement

Priority:

Version:

420+

Hardware:

Mac

OS:

OS X 10.4

Attachments:

Description	Flags
fix, test case and changelogs	hyatt: review-
new patch	none
patch, test case and change log	hyatt: review+

David Carson

Reported 2006-05-25 13:54:03 PDT

If the opacity of an object has a value of 0, we don't need to create a transparency layer and render the contents as it is not visible. Note that a block marked with opacity 0 is different to a block with visibility:none as visibility can be changed by an inner element to visibility:visible, whereas a inner can not change the restore the opacity to 1.

Attachments
fix, test case and changelogs (14.33 KB, patch) 2006-05-25 15:49 PDT, David Carson	hyatt: review-	Details Formatted Diff Diff
new patch (12.08 KB, patch) 2006-05-27 11:31 PDT, David Carson	no flags	Details Formatted Diff Diff
patch, test case and change log (11.94 KB, patch) 2006-05-30 18:26 PDT, David Carson	hyatt: review+	Details Formatted Diff Diff
Show Obsolete (2) View All Add attachment proposed patch, testcase, etc.

David Carson

Comment 1 2006-05-25 15:49:00 PDT

Created attachment 8545 [details] fix, test case and changelogs

Dave Hyatt

Comment 2 2006-05-26 12:02:47 PDT

Comment on attachment 8545 [details] fix, test case and changelogs Cool. Just a couple of comments. Please keep the name isTransparent. Should make the patch a lot smaller. :) Also fix the 2-space indenting. This showed up in a couple of spots in the patch.

Dave Hyatt

Comment 3 2006-05-26 13:11:05 PDT

There's another issue with this patch. It's not enough to stop your layer from painting. You also have to not paint any child layers either. The current patch will actually cause child layers to become visible.

David Carson

Comment 4 2006-05-27 11:31:42 PDT

Created attachment 8572 [details] new patch Modified patch to include suggestions from Hyatt.

Dave Hyatt

Comment 5 2006-05-29 23:01:49 PDT

Comment on attachment 8572 [details] new patch r=me Changelog comment should exclude the bit about renaming isTransparent. Indentation for the patch is still wrong (2-space instead of 4-space), but I won't minus again. Committer should make sure to fix up the indentation though.

David Carson

Comment 6 2006-05-30 18:26:18 PDT

Created attachment 8613 [details] patch, test case and change log removed tabs and removed extra comments from ChangeLog

Dave Hyatt

Comment 7 2006-05-30 18:40:45 PDT

Comment on attachment 8613 [details] patch, test case and change log r=me

Darin Adler

Comment 8 2006-06-02 08:55:33 PDT

Comment on attachment 8613 [details] patch, test case and change log I don't understand the relationship between this patch and these test cases. As far as I can tell, the patch simply optimizes drawing, and should have no effect on the test cases at all. I think it's good to have these new tests, but it's misleading to say that these test this bug. Is there a way to test to see if this this optimization is done?

Darin Adler

Comment 9 2006-06-02 09:48:21 PDT

David Carson caught up with me on IRC and clarified that these are test cases that check there's no regression from the change; they don't test the fix because it's a performance optimization.

Timothy Hatcher

Comment 10 2006-06-02 18:10:37 PDT

Landed in r14701.

Note You need to log in before you can comment on or make changes to this bug.