271047 - row group reflow triggers ###!!! ASSERTION: Computed overflow area must contain frame bounds

Status: RESOLVED FIXED

(Core :: Layout: Tables, defect)

Description

[Bernd]

load attahed testcase into a debug build
see the assertion the aReflowState.availableWidth contains a negative width,
which  leads to a overflow area with a negative width...

Comment 1

[Bernd]

Attachment: testcase

I tried to make the testcase strict and validating.., the validator as mozilla
does not like a script at this place

Comment 2

[timeless]

*** Bug 266753 has been marked as a duplicate of this bug. ***

Comment 3

[Bernd]

taking

Comment 4

[Bernd]

Attachment: patch

Comment 5

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

Am I missing something or does this guarantee that availSize.width is always *at
most* zero?

Comment 6

[Bernd]

yep thats the intention

Comment 7

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

You are guaranteeing that the available width/height is always negative or zero.
That can't be what you really want.

Comment 8

[Bernd]

My debugger tells me something different ;-).
The patch could be also translated into: don't substract more than you have
scenario a: availWidth > padding +spacing 
            substract padding and spacing
scenario b: availWidth < padding +spacing 
            substract only the availWidth and get a zero

I can of course also add a clean availWidth = PR_MAX(0, availWidth); if that is
more obvious.

Comment 9

[Bernd]

Attachment: more obvious patch

Comment 10

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

Comment on attachment 178442 [details] [diff] [review]
more obvious patch

that's much clearer :-)

Comment 11

[Bernd]

fix checked in