Bug 52708 - 3 CSS font tests are failing on commit-queue bots after 10.6.6 upgrade
Summary: 3 CSS font tests are failing on commit-queue bots after 10.6.6 upgrade
Status: NEW
Alias: None
Product: WebKit
Classification: Unclassified
Component: New Bugs (show other bugs)
Version: 528+ (Nightly build)
Hardware: PC OS X 10.5
: P2 Normal
Assignee: Nobody
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-01-19 02:52 PST by Eric Seidel (no email)
Modified: 2011-01-19 10:36 PST (History)
5 users (show)

See Also:


Attachments
failing layout test results from cr-jail-3 (cr-jail-4 fails the same way) (68.51 KB, application/zip)
2011-01-19 02:52 PST, Eric Seidel (no email)
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description Eric Seidel (no email) 2011-01-19 02:52:42 PST
Created attachment 79406 [details]
failing layout test results from cr-jail-3 (cr-jail-4 fails the same way)

3 CSS font tests are failing on commit-queue bots after 10.6.6 upgrade

These fonts don't fail on my local machine, only on these commit-queue bots.

It's very possible that the problem is with the bots and not our webkit code, but I'm creating a bug to have somewhere to record what we learn about the issue.

The failures were first noticed today.  The bots were upgraded to 10.6.6 yesterday.  It's also possible WebCore changes are causing the failures.

I've looked at the machine and the Helvetica fonts(which is the san-serif default font, which is what the test seems to be having trouble with) seem to match the ones on my (working) 10.6.6 desktop, at least in size.
Comment 1 Eric Seidel (no email) 2011-01-19 02:56:31 PST
(In reply to comment #0)
> These fonts don't fail on my local machine, only on these commit-queue bots.

I mean "these font tests" :)
Comment 2 Eric Seidel (no email) 2011-01-19 02:57:19 PST
If someone familiar with the font code might take a guess at what could be causing the attached failure results (see attached zip), that would be most helpful. :)
Comment 3 Eric Seidel (no email) 2011-01-19 03:30:16 PST
   .six {font: italic small-caps 100 150%/300% sans-serif;}
and
   .eleven {font: italic small-caps 100 150%/300% sans-serif;}

Are the lines of CSS corresponding to the failing text.

The text seems to be wrapping differently on the failing machines (as though the font widths are slightly different?)

This is all easy to see in the layout test results.  I've not tried to reduce the tests cases yet (mostly because that would involve multi-machine debugging...)