[Webkit-unassigned] [Bug 54517] Ensure loading has stopped in HistoryController::goToItem

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=54517





--- Comment #5 from Charles Reis <creis at chromium.org>  2011-02-16 12:04:05 PST ---
(In reply to comment #3)
> Oh, nevermind... I see that Page::goToItem special cases navigations that do not change the document!  Good :-)

Interesting point!  Page::goToItem actually only special cases state navigations, not fragment navigations.  As a result, we do end up stopping all loaders even if you're just going forward to a fragment.  This is visible in Safari, as well as in Chromium with my patch.  Repro steps:

1. Start at a simple page like about:blank.
2. Go to a page with an onload handler plus a slow, uncacheable resource (e.g., in an img tag).
3. Go to a fragment on that page.
4. Go back twice to the first page.
5. Click forward, then forward again before the onload handler fires.

The onload handler will never fire.

I'd argue that we should change Page::goToItem to avoid stopping all loaders for a fragment navigation, which would fix this in both Safari and this Chromium patch.  (Note that Page::goToItem already makes a similar exception for the database thread.)

And yes, this would require changing or removing the assert in HistoryController.

Sound reasonable?

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.