[Webkit-unassigned] [Bug 69296] REGRESSION (r95852?): Disappearing Border on bugs.webkit.org attachments <table>

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Wed Oct 5 09:46:57 PDT 2011


--- Comment #10 from Konstantin Shcheglov <scheglov at google.com>  2011-10-05 09:46:56 PST ---
(In reply to comment #9)
> (From update of attachment 109802 [details])
> This change doesn’t look right. Either it fixes a bug, so it needs a test case, or it doesn’t, so it needs an explanation for the rationale of the changes other than "fixes a detail".
> This makes CollapsedBorderValue objects larger, so there needs to be some benefit. From reading the comments in the bug report it seems that this *does* fix a bug and so needs a test case to show what it fixes.

1. Actual count of CollapsedBorderValue objects is small even for large tables, because only unique objects are remembered.

2. I think that my comment #7 has enough description why we need to store BorderValue as value. Is it not good enough? Is having description in bug comments not enough? I don't know what is policy about ChangeLog, so may be I should also include same description into ChangeLog file?

3. As you can see, this is memory problem, which happens _sometimes_, depending on whether we are "lucky" that m_style flag filled with 0, or not. So, it would be not easy to make reproducable test. Again, I'm open to recommendations here. If you know that in cases like this it is recommended to write at least flaky test, I will do this (hopefully eventSender.mouseMoveTo() works). I just want to describe reason why it was not done, and I'd like to have confirmation.

Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

More information about the webkit-unassigned mailing list