Bug 63591

Summary:

[Chromium] isRedirect is incorrectly determined for decidePolicyForNavigation

Product:

WebKit

Reporter:

Ben Smith <binji>

Component:

New Bugs

Assignee:

Nobody <webkit-unassigned>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

abarth, webkit.review.bot

Priority:

Version:

528+ (Nightly build)

Hardware:

Unspecified

OS:

Unspecified

Attachments:

Description	Flags
Patch to fix this bug.	none
Patch	none

Ben Smith

Reported 2011-06-28 18:15:33 PDT

Created attachment 99017 [details] Patch to fix this bug. In FrameLoaderClientImpl::dispatchDecidePolicyForNavigationAction() isRedirect is currently determined by hasRedirectChain(), which checks for an non-empty redirect chain. Instead, it should check for a redirect chain with more than one element. All valid redirects will have at least the source URL and the destination URL in the redirect chain.

Attachments
Patch to fix this bug. (1.30 KB, patch) 2011-06-28 18:15 PDT, Ben Smith	no flags	Details Formatted Diff Diff
Patch (4.57 KB, patch) 2011-08-30 15:40 PDT, Ben Smith	no flags	Details Formatted Diff Diff
Show Obsolete (1) View All Add attachment proposed patch, testcase, etc.

Ben Smith

Comment 1 2011-06-28 18:17:30 PDT

I don't believe a LayoutTest is necessary for this patch, because isRedirect is only used in chromium (for determining whether to fork a process), and has no user-visible effect in the browser.

Adam Barth

Comment 2 2011-06-29 11:48:05 PDT

Comment on attachment 99017 [details] Patch to fix this bug. All patch require a ChangeLog: http://www.webkit.org/coding/contributing.html

Adam Barth

Comment 3 2011-06-29 11:51:07 PDT

Comment on attachment 99017 [details] Patch to fix this bug. Is this testable with a Chromium WebKit API test? Generally, these sorts of subtle changes are easy to regress if we don't have any test coverage. If this is used for forking processes in the embedder, perhaps we can write a test at that layer? (Marking r- because of the lack of changelog.)

Ben Smith

Comment 4 2011-06-29 12:10:28 PDT

Yes, I believe it would be possible to test at that level. Is it WebKit/Source/WebKit/chromium/tests/*?

Ben Smith

Comment 5 2011-08-30 15:40:22 PDT

Created attachment 105710 [details] Patch

Ben Smith

Comment 6 2011-08-31 14:36:29 PDT

It's been a while since I've had time to work on this one... :) Adam, can you review my new patch? I've included a test in WebFrameTest that exhibits the bug (when the fix in FrameLoaderClientImpl.cpp is removed).

Adam Barth

Comment 7 2011-08-31 14:44:06 PDT

Comment on attachment 105710 [details] Patch Looks great. Thanks for adding the ChangeLog and test.

Ben Smith

Comment 8 2011-09-06 18:47:10 PDT

Adam, is there something else I need to do to commit this patch? It seems that it is in an infinite commit-queue. :)

Adam Barth

Comment 9 2011-09-06 19:33:53 PDT

Comment on attachment 105710 [details] Patch We just need to approve it for the queue!

WebKit Review Bot

Comment 10 2011-09-06 20:00:01 PDT

Comment on attachment 105710 [details] Patch Clearing flags on attachment: 105710 Committed r94630: <http://trac.webkit.org/changeset/94630>

WebKit Review Bot

Comment 11 2011-09-06 20:00:06 PDT

All reviewed patches have been landed. Closing bug.

Note You need to log in before you can comment on or make changes to this bug.