Bug 64895 – Wrong URL loaded into child frame after back/forward navigation

• Status: RESOLVED
• Resolution: FIXED
• Component: WebCore Misc.

[Nate Chapin]

Original report at http://code.google.com/p/chromium/issues/detail?id=59576.

Repro steps:
1. Display a web page
2. In the web page, display an iframe or object containing html
3. Navigate back & forward
4. In the same web page, display another iframe containing a different html page

The problem is in the logic in FrameLoader::loadURLIntoChildFrame to load a child frame from a HistoryItem instead of from the url.  I think we only want to load from a HistoryItem before the load event.

Note that this problem is prevented by the page cache because the old contents of the child frame are immediately present and the display of a  new frame will typically add a new Frame rather than overwrite.

[Nate Chapin]

Created attachment 101509 [details]
patch

[Mihai Parparita]

Comment on attachment 101509 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=101509&action=review

Is the current behavior (that the URL is restored for iframes for back-forward navigations) tested anywhere?

> LayoutTests/fast/loader/child-frame-add-after-back-forward.html:9
> +    layoutTestController.overridePreference('WebKitUsesPageCachePreferenceKey', 0);

You may want to add an unload handler too, so that the page cache is disabled when running the test in Safari too.

[Nate Chapin]

(In reply to comment #2)
> (From update of attachment 101509 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=101509&action=review
> 
> Is the current behavior (that the URL is restored for iframes for back-forward navigations) tested anywhere?

Yeah, if I comment out the if() statement and everything inside, 2 tests each in fast/history and http/tests/navigation fail.

> 
> > LayoutTests/fast/loader/child-frame-add-after-back-forward.html:9
> > +    layoutTestController.overridePreference('WebKitUsesPageCachePreferenceKey', 0);
> 
> You may want to add an unload handler too, so that the page cache is disabled when running the test in Safari too.

Good point.

[Nate Chapin]

Created attachment 101743 [details]
Patch for landing

[WebKit Review Bot]

Comment on attachment 101743 [details]
Patch for landing

Clearing flags on attachment: 101743

Committed r91583: <http://trac.webkit.org/changeset/91583>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.

[Alexey Proskuryakov]

*** Bug 64402 has been marked as a duplicate of this bug. ***