Bug 45576 - fast/files/apply-blob-url-to-img.html timeout on Leopard
Summary: fast/files/apply-blob-url-to-img.html timeout on Leopard
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: WebCore JavaScript (show other bugs)
Version: 528+ (Nightly build)
Hardware: All OS X 10.5
: P2 Normal
Assignee: Jian Li
URL:
Keywords:
Depends on: 45802
Blocks:
  Show dependency treegraph
 
Reported: 2010-09-10 16:37 PDT by Jian Li
Modified: 2012-02-16 11:40 PST (History)
7 users (show)

See Also:


Attachments
Proposed Patch (8.02 KB, patch)
2010-09-10 17:05 PDT, Jian Li
levin: review+
jianli: commit-queue-
Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Jian Li 2010-09-10 16:37:53 PDT
fast/files/apply-blob-url-to-img.html timeout on Leopard.
Comment 1 Jian Li 2010-09-10 17:05:19 PDT
Created attachment 67267 [details]
Proposed Patch
Comment 2 David Levin 2010-09-14 14:43:12 PDT
Comment on attachment 67267 [details]
Proposed Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=67267&action=prettypatch

> WebKit/chromium/ChangeLog:8
> +        * src/ResourceHandle.cpp: Remove the unneeded static comment.
s/unneeded/no longer applicable/
Comment 3 Jian Li 2010-09-14 15:27:58 PDT
Committed as http://trac.webkit.org/changeset/67503.
Comment 4 WebKit Review Bot 2010-09-14 16:04:14 PDT
http://trac.webkit.org/changeset/67503 might have broken GTK Linux 32-bit Release
Comment 5 Alexey Proskuryakov 2010-09-14 16:08:08 PDT
I don't think that this is a good fix. supportsBufferedData is a check for platform capabilities, and adding a check that internal to WebCore on top of it is highly confusing. I think that a different fix is needed here.

Also, I may not understand the purpose of BlobResourceHandle, but it seems to share the same problem - although ResourceHandle is supposed to be a platform abstraction, it now has a subclass that contains WebCore logic.
Comment 6 Alexey Proskuryakov 2010-09-14 16:08:36 PDT
Since this also broke build, rolling out seems like an appropriate course of action to me.
Comment 7 Alexey Proskuryakov 2012-02-16 10:43:46 PST
This change didn't unskip the test. Done in <http://trac.webkit.org/changeset/107879>.

It is disappointing that the above comments have not be acted upon.
Comment 8 Jian Li 2012-02-16 11:40:06 PST
(In reply to comment #7)
> This change didn't unskip the test. Done in <http://trac.webkit.org/changeset/107879>.
> 
> It is disappointing that the above comments have not be acted upon.

My apology for not addressing the concern quickly. We did go through quite a lot discussions on which layer we wanted to hook up with the URL request processing. I agree that BlobResourceHandle inherited from ResourceHandle might not be a ideal solution to process the blob URL request. But the discussions did not lead to any other better choice at that time.