Open Bug 566159 Opened 14 years ago Updated 9 months ago

ASSERTION: should have done initial reflow by now in docshell/base/crashtests/436900-2.html and docshell/base/crashtests/1257730-1.html

Categories

(Core :: DOM: Navigation, defect)

defect

Tracking

()

People

(Reporter: tnikkel, Unassigned)

References

Details

In addition to that assertion there are some other problems.

If you open docshell/base/crashtests/436900-2-inner.html#foo (or 436900-1-inner.html) directly then the tab title will have "436900-2-inner.html#foo+++" but the location bar will have "436900-2-inner.html#foo".

If you open docshell/base/crashtests/436900-2-inner.html (or 436900-1-inner.html) directly then the browser will hang indefinitely.

(There is a typo in the test file name, it is actually for bug 436990.)
Is this a regression?  The assert in the summary isn't a docshell bug offhand, is it?
Not a regression that I know of. I added the assert recently.

The docshell code is calling PresShell::ScrollToAnchor before we have any frames, so it will fail, whether you call that a docshell bug or a bug somewhere else I guess doesn't matter.

The other two issues are probably fairly classified as docshell.
> The docshell code is calling PresShell::ScrollToAnchor before we have any
> frames, so it will fail

OK.  And the point is that nothing will retry the scroll, right?

Perhaps the scroll retry needs to move into the presshell, then?  I don't see how anyone else can handle it correctly...
Yeah, PresShell::DoScrollContentIntoView sets mContentToScrollTo to null if it doesn't have a frame, thus ending any attempt to scroll to it.

That sounds like a good plan.

As part of work to Fission-ize the reftest harness (bug 1561766) we encounter another instance of this assert in docshell/base/crashtests/1257730-1.html

The reason that this new instance shows up is that the Fission work on the reftest harness, in general, makes things more async. So in this docshell/base/crashtests/1257730-1.html we previously tore down the page and moved on to the next test before being able to hit these assertions.

The asserts are slightly different because the stacks in docshell/base/crashtests/436900-2.html come from being asked to scroll, for docshell/base/crashtests/1257730-1.html the stacks come from a focus call from an autofocus form element.

I plan to annotate the new asserts as expected and point to this bug.

Summary: ASSERTION: should have done initial reflow by now in docshell/base/crashtests/436900-2.html → ASSERTION: should have done initial reflow by now in docshell/base/crashtests/436900-2.html and docshell/base/crashtests/1257730-1.html
Severity: normal → S3
See Also: → 1708332
You need to log in before you can comment on or make changes to this bug.