131020 - 100% height IFRAME and div badly sized in nested %-height TABLEs

Status: VERIFIED FIXED

(Core :: Layout: Tables, defect, P2)

Description

[Kai Lahmann (is there, where MNG is)]

just look at that URl - the <iframe> is correct (Validator gives no Error), but
Mozilla shows nothing

Comment 1

[Mats Palmgren (inactive)]

Attachment: Testcase

Comment 2

[Mats Palmgren (inactive)]

Confirming bug, 2002-03-13-03 trunk on Windows 98 SE.

Comment 3

[Kevin Ar18]

Is this a duplicate of Bug 41656 ?
http://bugzilla.mozilla.org/show_bug.cgi?id=41656

Comment 4

[Bernd]

Attachment: testcase strict with a div

Comment 5

[Bernd]

looks more like a table issue as the same problem happens with a div.

Comment 6

[Bernd]

Attachment: another testcase

this testcase does not render correctly with normal build but works with viewer
(debug build).

Comment 7

[Bernd]

Attachment: pure table testcase

Comment 8

[Bernd]

the testcase has rendered correctly in 2000111004 WINNT

Comment 9

[karnaze (gone)]

*** Bug 133948 has been marked as a duplicate of this bug. ***

Comment 10

[Kevin McCluskey (gone)]

nsbeta1+

Comment 11

[karnaze (gone)]

Attachment: patch to fix the bug

Comment 12

[karnaze (gone)]

Attachment: some more test cases

Comment 13

[Bernd]

The special height reflow how does it relate to the conventional reflow's
can it be a initial or a style change reflow or is that a special type of 
incremental reflow subtype resize?
Why a special height reflow can be a premature reflow? When and why is it 
a mature reflow?
What happens for the scenario in bug 115245 a table cell containing another 
table with height 100% and and below that a second table. As soon as the table 
cell has a height_cell != unconstrained_size first table will grow to the 
heigt_cell and the second table has it's own height t2_height. How does the 
table cell finally arrive at height_cell+t2_height?

Some question about the concept: 
1. we have frame who has a percent height, this will be the
mPercentHeightReflowInitiator
2. sometimes other frames like a div block can also have a percent height do
they also set 
   that variable? If not, why dont they set this variable.
3. will we end at a reflow per percent frame?
   <table style="height:100%">
     <tbody style="height:100%">
       <tr style="height:100%">
         <td style="height:100%">
   how many reflows will be necessary to reflow this one, or who will be the 
   mPercentHeightReflowInitiator in this case?
4. How does this concept work together with the PercentHeightObserver
5. What does  this PercentHeightObserver do?
6. Why did you remove class nsIPercentHeightObserver; from nsIFrame.h?
7. What happens if the content has a larger maxelementsize.height then the
computed.height?
 

just a offtopic: why does viewer ( my favourite application in this browser suite) 
issue more reflows? Does that mean to debug I need to start the bloated monster??

Comment 14

[karnaze (gone)]

Bernd, I'll be attaching a new patch with comments addressing most of your 
questions. For the ones that aren't covered.

>>6. Why did you remove class nsIPercentHeightObserver; from nsIFrame.h?
it wasn't needed

>>7. What happens if the content has a larger maxelementsize.height then the
>>computed.height?
max element height isn't used anywhere in layout
 
>>just a offtopic: why does viewer ( my favourite application in this browser 
>>suite) issue more reflows? 
I don't have time to figure out why.

>>Does that mean to debug I need to start the bloated >>monster??
I guess, if you want to be 100% sure

Comment 15

[karnaze (gone)]

Attachment: revised patch with reviewer's suggestions

Comment 16

[Bernd]

Comment on attachment 76783 [details] [diff] [review]
revised patch with reviewer's suggestions

r=bernd

Comment 17

[Marc Attinasi]

Comment on attachment 76783 [details] [diff] [review]
revised patch with reviewer's suggestions

sr=attinasi

Comment 18

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Comment on attachment 76783 [details] [diff] [review]
revised patch with reviewer's suggestions

a=roc+moz

Comment 19

[karnaze (gone)]

The patch is in.

Comment 20

[karnaze (gone)]

*** Bug 125267 has been marked as a duplicate of this bug. ***

Comment 21

[Ralf Vitasek]

nested 100% height-tables tables still don't shrink correctly in height on 
downsizing of the browser window (tested on windows nightly build 20020401)

Comment 22

[Bernd]

the testcases have been fixed, but the dupe relationship was wrong.

Comment 23

[Kevin Ar18]

Attachment: iframe within div

This is similar to some other testcases.
See testcase 75860 and tescase 75863.

Shows problem with iframes nested in a div.
height="100%" on iframes still does not work in this case

Comment 24

[Kevin Ar18]

Tested with latest Win32 nightly build (20022040303) and it works now for 
iframes nested withing tables.  However, it still does not work for iframes 
nested within div tags.

Comment 25

[karnaze (gone)]

That last test case is a different bug (and there may already be one filed). The 
outer table is getting no height because it is 100% the height of a div without 
a height. In standard mode this is correct, but it is not compatible with IE.  

Comment 26

[Kevin Ar18]

Can you confirm whether a new bug report has been reported on this issue?  Or 
should somebody just report a new one and see if it get's confirmed as a 
duplicate?

I did a quick search and didn't find anything, but then again I may have done 
it wrong.

Comment 27

[Kevin Ar18]

Well, I filed a new bug report.
It's under Bug 135706.

Comment 28

[Amarendra Hanumanula]

 All the testcases with the original problem works fine. Making verified.