163800 - Browser crash when close print preview in a page has fixed DIV

Status: RESOLVED WORKSFORME

(Core :: Print Preview, defect)

Description

[Pete Zha]

This bug is different with bug 156982. I will upload a testcase for this.
The testcase include a fixed DIV element. Please make sure your window's size is
small enough to see the scrollbar of the DIV before print preview.

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.1b) Gecko/20020821
Reproduce step: (Always if the window is small enough to see the scrollbar of
the fixed DIV element)
1) Open the first attachment
2 [details] [diff] [review]) Click print preview
3) Click "Close"

Browser will crash.

Comment 1

[Pete Zha]

Attachment: testcase

Comment 2

[Pete Zha]

Attachment: patch

Fix the crash.

Comment 3

[Pete Zha]

Comment on attachment 96134 [details] [diff] [review]
patch

rods, carry your r= in bug 156982 to have this one to be fixed quickly. Is it
OK?

Comment 4

[Johnny Stenback (:jst)]

Comment on attachment 96134 [details] [diff] [review]
patch

sr=jst

Comment 5

[Pete Zha]

checked in trunk

Comment 6

[Pete Zha]

checked in NETSCAPE_7_0_OEM_BRANCH (a=jdunn)

Comment 7

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

This caused a fairly serious regression in bug 163800.

You can't just skip the AttributeChanged logic if there are no scrollbars. The
problem is that if the content shrinks so that the scrollbars are not needed and
disappear, the AttributeChanged logic is required to scroll what's left of the
document back into view. Otherwise we can be left looking below the bottom of
the document with no way to scroll back up!

Please find out why this crashes print preview and fix that rather than whacking
layout here.

Comment 8

[Andrew Schultz]

> This caused a fairly serious regression in...

bug 168170

Comment 9

[Pete Zha]

ok, let's investigate it.

Comment 10

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

I think we should start by backing out the first part of the patch on the trunk,
to undo the regression.

Comment 11

[Pete Zha]

Attachment: patch to backout

backout first part of previou patch to resolve the layout regression.

Comment 12

[rods (gone)]

Comment on attachment 100360 [details] [diff] [review]
patch to backout

r=rods

Comment 13

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

Comment on attachment 100360 [details] [diff] [review]
patch to backout

sr=roc+moz

Comment 14

[Pete Zha]

Comment on attachment 100360 [details] [diff] [review]
patch to backout

checked in this backout patch

Comment 15

[Pete Zha]

see also bug 135653

Comment 16

[Olivier Cahagne]

shall we close this one ? Every bug depending on this one seem to have been fixed.

Comment 17

[Pete Zha]

sure, already checked in on comment 14

Comment 18

[Pete Zha]

oh, sorry, I should not close this bug! on comments 14, I just checked in the
backout patch. We should test if the orignial problem still exists.
->REOPEN

Comment 19

[Bill Mason]

The testcase WFM in 2003040105 PC/WinXP.  Probably this issue has resolved itself.

A similiar crash with fixed DIVs is occurring in bug 200774, FWIW.

Comment 20

[Laurin Rötzer]

Still/again there in 1.6final?
Try it at www.winfuture.de.
But at pages like this, it works well.

Comment 21

[James Napolitano]

In Mozilla 1.7RC2 on Win98se, both the testcase and http://www.winfuture.de/ do
not crash for me on print preview.  Is this bug still present?

Comment 22

[Laurin Rötzer]

http://www.winfuture.de still crashes with 1.7RC2 (clean install on WinXP)

Comment 23

[Martijn Wargers (dead)]

Is this bug still present in builds at/after 2005-03-23?

Comment 24

[Wayne Mery (:wsmwk)]

WFM no crash with testcase or http://www.winfuture.de/

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9a1) Gecko/20050907 Firefox/1.6a1
Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9a1) Gecko/20051020 SeaMonkey/1.1a

and also on XP - recent trunk builds but not same as above

Comment 25

[Wayne Mery (:wsmwk)]

-> WFM