Bug 76590 - REGRESSION (r104060): Page contents not painted if inserting a new stylesheet and temporary body node
Summary: REGRESSION (r104060): Page contents not painted if inserting a new stylesheet...
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: CSS (show other bugs)
Version: 528+ (Nightly build)
Hardware: All All
: P2 Normal
Assignee: Antti Koivisto
URL: http://persistent.info/webkit/test-ca...
Keywords: HasReduction, Regression
Depends on:
Blocks:
 
Reported: 2012-01-18 17:10 PST by Mihai Parparita
Modified: 2012-03-02 02:47 PST (History)
3 users (show)

See Also:


Attachments
patch (9.60 KB, patch)
2012-03-02 02:07 PST, Antti Koivisto
mjs: review+
Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Mihai Parparita 2012-01-18 17:10:35 PST
This was originally filed as http://crbug.com/109888, see http://persistent.info/webkit/test-cases/crbug-109888.html for a reduction (the text "The body contents should be visible" should appear on a green background, instead the entire page is white).

The appending of the stylesheet dynamically is something that the page did, the inserting of the (temporary) body tag is from jQuery pre-1.6.2 which was included in the page (originally via a Chrome extension, though that's not relevant, since it happens when done directly too).
Comment 1 Antti Koivisto 2012-01-19 04:41:38 PST
Delightfully obscure.
Comment 2 Mihai Parparita 2012-02-04 18:30:13 PST
Another regression from this change is http://stationstreetpgh.com/ (it should have pie charts that animate in, but they don't appear at all in WebKit nightlies post-r104060)

So far, it seems to boil down to the "active" callback in Typekit's font loading code not firing (I've asked someone from their side if they can debug it).

Here's a somewhat reduced test case: http://persistent.info/chromium/test-cases/crbug-112714/page.html

It should display "loading" and then "active". In builds with the bug, it just displays "loading", and then eventually "inactive" (presumably a failure timeout firing).
Comment 3 Antti Koivisto 2012-03-02 02:07:57 PST
Created attachment 129854 [details]
patch
Comment 4 Maciej Stachowiak 2012-03-02 02:37:10 PST
Comment on attachment 129854 [details]
patch

r=,e
Comment 5 Antti Koivisto 2012-03-02 02:47:29 PST
trac.webkit.org/changeset/109543