136441 - <select> dropdowns messed up if they are in an iframe and the dropdown is longer than the iframe's height

Status: VERIFIED FIXED

(Core :: Layout: Form Controls, defect)

Description

[Doron Rosenberg (IBM)]

See the top right dropdown at http://www.sueddeutsche.de.  The part of the
dropdown below the iframe height (see http://www.sueddeutsche.de/sz/head.php for
the iframe's src) is rendered as a brown box.

Also seeing with XBL form controls on.

Not that this will help, but this happens at a top 100 .de Evangelism site
(bugscape 13585)

Comment 1

[Robert Kaiser]

Doron saw this on WinXP, I'm seeing this on Linux.
It looks like the dropdown popup's contents are simply not painted if they're
outside the <iframe> borders. The shown items are clickable and working, the not
painted part is also not reacting to any mouse clicks. Using arrow keys on the
keyboard, you can navigate through the popup's contents, though (still not painted).
Interestingly, it works fine if I show the document (head.php) in a new window
(no iframe) and resize the moz window so that the dropdown goes outside the
window. It's painted fine then, and working.

Comment 2

[Doron Rosenberg (IBM)]

btw, this works in Netscape 6.2.2, so a regression.

Comment 3

[Mattias Holmlund]

This also happens on www.teletrade.se (when logged in), a Swedish online broker.
RC1-build, WinNT and Win98. It is possible to select values in the dropdown by
using the arrow-keys, but it's very ugly.

Comment 4

[Doron Rosenberg (IBM)]

-> jkeiser as per irc

Comment 5

[Arthur]

Adding regression keyword based on comment #2.

Comment 6

[John Keiser (jkeiser)]

Roc, this seems suspiciously like a bug in the new view manager.

Comment 7

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

Yes, it does sound like a view manager problem. I'll steal it for now.

Comment 8

[Christian Schmidt]

*** Bug 146393 has been marked as a duplicate of this bug. ***

Comment 9

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

Attachment: Testcase part 1

Comment 10

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

Attachment: Minimal testcase demonstrating problem with clipped IFRAMES

The problem seems to be with an IFRAME that is clipped; the clipping is being
applied to the popup.

Comment 11

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

Attachment: the fix

Here's the fix. This patch does two things:
1) Removes GetFloating from the public nsIView interface and deXPCOMifies it
2) Makes a simple, small behavior change in nsView::GetClippedRect to fix this
bug: as we go up the view tree searching for clippers, stop if we transition
out of a floating view to a non-floating parent.

Comment 12

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

Need r/sr from kmcclusk+waterson

Comment 13

[Doron Rosenberg (IBM)]

since this breaks a top german site, how would you feel with this landing for
1.0.1 regarding any possible regressions?

Comment 14

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

I'd be comfortable with that.

Comment 15

[Doron Rosenberg (IBM)]

Nominating for nsbeta1, since it breaks a german top newspaper site (see
http://bugscape.mcom.com/show_bug.cgi?id=13585).  Iframes are pretty popular in
europe and especially so in Asia...

Comment 16

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

Let me give you a minimal patch suitable for the branch (i.e., without the cleanup).

Comment 17

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

Attachment: Minimal patch

Here you go. We should try to get this into trunk and branch.

Comment 18

[Chris Waterson]

Comment on attachment 85238 [details] [diff] [review]
Minimal patch

sr=waterson

Comment 19

[Kevin McCluskey (gone)]

Comment on attachment 85238 [details] [diff] [review]
Minimal patch

r=[e-mail]

Comment 20

[Doron Rosenberg (IBM)]

nominating for 1.0.1 as well.  Do you want to email drivers or should I?

Comment 21

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

Let me check into the trunk first (should happen tonight or tomorrow). Once
that's done and it's baked for a few days, you can email drivers (I always hate
mailing drivers since I am one :-).)

Comment 22

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

Fix checked in on trunk.

Comment 23

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

Note that if this is checked into the branch then you should also check in the
fix for bug http://bugzilla.mozilla.org/show_bug.cgi?id=148350, a compiler
optimization bug.

Comment 24

[Randell Jesup [:jesup] (needinfo me)]

Comment on attachment 85238 [details] [diff] [review]
Minimal patch

please check into the 1.0.1 branch ASAP. once landed remove the 
mozilla1.0.1+ keyword and add the fixed1.0.1 keyword

Comment 25

[Randell Jesup [:jesup] (needinfo me)]

Please make sure the patch for bug 148350 gets checked in with this

Comment 26

[Matthias Versen [:Matti]]

*** Bug 150505 has been marked as a duplicate of this bug. ***

Comment 27

[Doron Rosenberg (IBM)]

Here is a combinded patch for this and the dependant compile bug that applies
and builds on the branch:

http://nexgenmedia.net/finalroc.txt (has roc's approval)

Now someone with cvs access needs to check it in :)

Comment 28

[Terri Preston]

This is indeed fixed on win xp branch build 2002062708 and mac OS X branch build
2002062705 BUT is not fixed linux branch build 2002062606, removing fixed1.0.1
keyword and changing os to linux

Comment 29

[Terri Preston]

Verified fixed win XP trunk build 2002062713, Mac OS Trunk build 2002062712 and
linux trunk build 2002062713, so marking verified (only open issue in this bug
then is linux branch)

Comment 30

[Terri Preston]

Verified linux branch build 2002082207

Comment 31

[Michael Lefevre]

*** Bug 143337 has been marked as a duplicate of this bug. ***