Bug 48168

Summary:

WebP decoding hits ASSERT

Product:

WebKit

Reporter:

Adam Barth <abarth>

Component:

New Bugs

Assignee:

Adam Barth <abarth>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

abarth, commit-queue, pascal.massimino, pkasting

Priority:

Version:

528+ (Nightly build)

Hardware:

Other

OS:

OS X 10.5

Attachments:

Description	Flags
Patch	none

Description Adam Barth 2010-10-22 18:12:19 PDT

WebP decoding hits ASSERT

Comment 1 Adam Barth 2010-10-22 18:19:46 PDT

Created attachment 71616 [details]
Patch

Comment 2 Andreas Kling 2010-10-23 03:38:00 PDT

Comment on attachment 71616 [details]
Patch

LGTM.

Comment 3 Adam Barth 2010-10-23 09:34:13 PDT

Comment on attachment 71616 [details]
Patch

Thanks.

Comment 4 WebKit Commit Bot 2010-10-23 09:48:14 PDT

Comment on attachment 71616 [details]
Patch

Clearing flags on attachment: 71616

Committed r70392: <http://trac.webkit.org/changeset/70392>

Comment 5 WebKit Commit Bot 2010-10-23 09:48:19 PDT

All reviewed patches have been landed.  Closing bug.

Comment 6 Pascal Massimino 2010-10-23 11:35:10 PDT

(In reply to comment #5)
> All reviewed patches have been landed.  Closing bug.

thanks for fixing that Adam!
i'm a little worried that rgb.resize() is zero'ing the memory we are going to be
filling. Is it? If so, this is unneeded operation... But we can revisit that later.

Comment 7 Adam Barth 2010-10-23 17:19:56 PDT

> i'm a little worried that rgb.resize() is zero'ing the memory we are going to be
> filling. Is it? If so, this is unneeded operation... But we can revisit that later.

We should look at the code an see.  There area a bunch of templates to unwind.  If it does, we can add a resizeUninitialized() function to vector.