304378 - Parsing incorrect number of frameset rows or cols

Status: RESOLVED INVALID

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

Description

[Barrie Lloyd-Jones]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; pt-BR; rv:1.7.8) Gecko/20050511 Firefox/1.0.4
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; pt-BR; rv:1.7.8) Gecko/20050511 Firefox/1.0.4

please check site http://www.mj.gov.br/sistemas/EstrangeiroWEB/index.asp
This site opens and operates normally with Internet Explorer but only partially
with Firefox. I want to standardize Forefox in my company but due to this glich
I am obliged to continue using IE at least for this page.
Please help me find a way to adjust my configs as necessary.
Your help is much appreciated
Thanks
Barrie

Reproducible: Always

Steps to Reproduce:
1.just go to the link, page opens but is incomplete, search field do not appear
2.
3.

Actual Results:  
page opens partcially and stops

Comment 1

[Phil Ringnalda (:philor)]

While I suspect the answer will be either "invalid" or "tech evangelism", it
does at first glance look like IE's error recovery might be simple enough to
emulate.

The problem is that someone commented out a random chunk of the page, so the
outer frameset says rows="131,*" but there's only one row, not two, so we only
display the one col that the inner frameset says to display.

From just a little testing, it looks like what IE does is copy extra values down
to children, if they aren't specified there, treating

<frameset rows="131,*">
  <frameset cols="*">
    <frame>
    <frame>
  </frameset>
</frameset>

as being

<frameset rows="131,*" cols="*">
  <frame>
  <frame>
</frameset>

No idea how many pages depend on that behavior, or how many depend on our
alternate behavior, though.

Comment 2

[Phil Ringnalda (:philor)]

Attachment: testcase

We see this as one row of 131 (though we actually load google.com, despite not
having anywhere to show it), IE6 sees it as two rows and one col.

Comment 3

[Boris Zbarsky [:bzbarsky]]

This is invalid; given the ways people do nested framesets, IE's behavior is almost certainly rather more complicated than comment 1 suggests, and reverse-engineering it just isn't worth it.