Closed Bug 430050 Opened 17 years ago Closed 17 years ago

xul:browser changes break Firebug

Categories

(Core :: DOM: Core & HTML, defect)

x86
All
defect
Not set
normal

Tracking

()

RESOLVED FIXED

People

(Reporter: smaug, Assigned: smaug)

References

Details

Attachments

(3 files)

Attachment #316804 - Flags: review?(jonas)
Gotta block, and we need some test coverage here, clearly. :/
Flags: blocking1.9+
yup, I'm writing tests.
Assignee: nobody → Olli.Pettay
I don't want to discourage your work on this problem, but it need not block 411814, firebug no longer uses loadURI in its own load sequence.
I would be extremely surprised if this didn't also affect other extensions, but yeah, thanks a ton for that parallel fix.
Attached patch a testSplinter Review
This is not a perfect test. I should still figure out how to test the behavior Firebug does. But useful test anyway.
Comment on attachment 316804 [details] [diff] [review] cancel frameloader initialization when needed. >+ // Don't use a temporary array for mInitializableFrameLoaders, because >+ // loading a frame may cause some other frameloader to be removed from the >+ // array. But be careful to keep the loader alive when starting the load! >+ while (mInitializableFrameLoaders.Length()) { >+ nsRefPtr<nsFrameLoader> loader = mInitializableFrameLoaders[0]; >+ mInitializableFrameLoaders.RemoveElementAt(0); >+ if (loader) { >+ loader->ReallyStartLoading(); Is the nullcheck really needed? >@@ -6808,16 +6808,28 @@ nsDocShell::InternalLoad(nsIURI * aURI, > return NS_ERROR_FAILURE; > } > > rv = CheckLoadingPermissions(); > if (NS_FAILED(rv)) { > return rv; > } > >+ // If this docshell is owned by a frameloader, make sure to cancel >+ // possible frameloader initialization before loading a new page. >+ nsCOMPtr<nsIDocShellTreeItem> parent; >+ GetParent(getter_AddRefs(parent)); >+ if (parent) { >+ nsCOMPtr<nsIDOMDocument> domDoc = do_GetInterface(parent); >+ nsCOMPtr<nsIDocument> doc = do_QueryInterface(domDoc); Silly that docshells don't GetInterface to nsIDocument :( r=me
Attachment #316804 - Flags: review?(jonas) → review+
Comment on attachment 316804 [details] [diff] [review] cancel frameloader initialization when needed. nullcheck is not needed. Doesn't cause any harm though.
Attachment #316804 - Flags: superreview?(jst)
Attachment #316804 - Flags: superreview?(jst) → superreview+
Attachment #316804 - Flags: approval1.9?
Comment on attachment 316804 [details] [diff] [review] cancel frameloader initialization when needed. a1.9+=damons
Attachment #316804 - Flags: approval1.9? → approval1.9+
No longer blocks: 411814
Status: NEW → RESOLVED
Closed: 17 years ago
Resolution: --- → FIXED
Component: DOM: Core → DOM: Core & HTML
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: