49874 - iframes and frames require reflow to load their documents

Status: RESOLVED FIXED

(Core :: Layout: Images, Video, and HTML Frames, defect, P3)

Description

[Brendan Eich [:brendan]]

This has led to bugs with onload firing to early, and is complicating other
modules such as XBL.  Asynchronous reflow makes things even more complicated. 
It might be good to know *why* only frames require a reflow to load their docs,
and not top-level windows, but hyatt points out that this glaring special case
predates the introduction of asynchronous reflow.

Waterson, we think you are the man to fix this one. Or otherwise to make us all
happy and blissful.  Don't make us cry now.

/be

Comment 1

[Chris Waterson]

Talked with brendan & hyatt about this yesterday. When the IFRAME code was
originally written, *all* reflows were done synchronously. Once we added
asynchronous reflow, we violated that assumption.

I think that this could cause ONLOAD problems for documents with IFRAMES,
specifically, the following race:

- IFRAME appears in last content batch, frame is constructed but not flowed.
- Reflow is enqueued, but doesn't fire because reflow timer suppresses it.
- outer URL completes, removes self from load group, active URL count goes to
zero
- outer doc's ONLOAD handler fires
- Reflow timer fires, causing IFRAME to be reflowed, kicking off inner load

There are two ways to fix this:

1. Move IFRAME docshell initialization and URL loading to frame constructor.
(Issues with zero w&h docshell?)

2. Force synchronous reflow when IFRAME is detected.

I believe that (1) would be the more elegant solution, but am not sure if there
are issues? (Images kick of load during frame construction, right?)

Anyway, this is not likely to be a huge issue, so futuring for now.

Comment 2

[Amarendra Hanumanula]

QA Contact update

Comment 3

[David Hyatt]

Bring it.  I'm fixing this NOW. ;)

Comment 4

[David Hyatt]

Attachment: Patch to fix this bug as well as 98109

Comment 5

[David Hyatt]

Ready for r/sr.

Comment 6

[jag (Peter Annema)]

Comment on attachment 49663 [details] [diff] [review]
Patch to fix this bug as well as 98109

r=jag

Comment 7

[David Hyatt]

Fixed.

Comment 8

[Boris Zbarsky [:bzbarsky]]

*** Bug 100502 has been marked as a duplicate of this bug. ***