289431 - Horizontal scrollbar appears periodically in Download Manager

Status: RESOLVED FIXED

(Core :: Layout, defect)

Description

[Hideo Oshima]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8b2) Gecko/20050404 Firefox/1.0+
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8b2) Gecko/20050404 Firefox/1.0+

In the case that the size of download file is unknown and Download Manager
window is big enough, Horizontal scroolbar appears periodically in Download
Manager while downloading the file.

Reproducible: Always

Steps to Reproduce:
1.Open Download Manager
2.Maximize Dowload Manager window
3.Click above URL(start downloading)

Actual Results:  
Horizontal scroolbar appears periodically in Download Manager.

Expected Results:  
Horizontal scrollbar doesn't appear in Download Manager.

Comment 1

[Hideo Oshima]

Typo in the summary. Sorry for SPAM.

Comment 2

[Martijn Wargers (dead)]

Attachment: Testcase

A testcase that shows the situation here. My connection is just too fast for
the url ;)

Comment 3

[Martijn Wargers (dead)]

Well, this could quite easily be fixed here:
http://lxr.mozilla.org/seamonkey/source/toolkit/content/widgets/progressmeter.xml#60
And here:
http://lxr.mozilla.org/seamonkey/source/xpfe/global/resources/content/bindings/progressmeter.xml#60
by not using overflow:-moz-hidden-unscrollable, but by using overflow:hidden; or
overflow:-moz-scrollbars-none; (Boris, is there a difference between the two?)

But isn't the underlying problem here that -moz-hidden-unscrollable isn't doing
the same as hidden/-moz-scrollbars-none? Should -moz-hidden-unscrollable not
behave the same here as hidden/-moz-scrollbars-none here?

Comment 4

[Boris Zbarsky [:bzbarsky]]

> (Boris, is there a difference between the two?)

No, except the latter is deprecated.

> Should -moz-hidden-unscrollable not behave the same here as
> hidden/-moz-scrollbars-none here?

"good question".  Probably, but roc should know better.

Comment 5

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

The problem is that overflow:-moz-hidden-unscrollable isn't turning off the
overflow area for the frame, so the outer scrollable box sees the overflow of
the progress-meter's children and thinks it needs a scrollbar.

I'll make a patch to fix that.

Comment 6

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

Attachment: fix

Fairly easy fix ... not sure who should review this

Comment 7

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 180340 [details] [diff] [review]
fix

>Index: layout/generic/nsFrame.cpp
>+  if (geometricOverflow &&
>+      GetStyleDisplay()->mOverflowX == NS_STYLE_OVERFLOW_CLIP) {

Assert in here that mOverflowY == CLIP if and only if mOverflowX == CLIP?  Just
in case that ever changes, for some odd reason?

>+nsBox::DoesClipChildren()
>+{
>+  return GetStyleDisplay()->mOverflowX == NS_STYLE_OVERFLOW_CLIP;

Similar here:

NS_ASSERTION((GetStyleDisplay()->mOverflowY == NS_STYLE_OVERFLOW_CLIP) ==
	     (GetStyleDisplay()->mOverflowX == NS_STYLE_OVERFLOW_CLIP),
	     "If one overflow is clip, the other should be too");

r+sr=bzbarsky with that.

Comment 8

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

Comment on attachment 180340 [details] [diff] [review]
fix

a fairly safe fix to a core layout issue that has UI polish repercussions

Comment 9

[Brendan Eich [:brendan]]

Comment on attachment 180340 [details] [diff] [review]
fix

a=me for 1.8b2.

/be

Comment 10

[Martijn Wargers (dead)]

Sometimes I browse through Mozilla's source code and I notice a thing:
http://bonsai.mozilla.org/cvsblame.cgi?file=mozilla/layout/style/nsRuleNode.cpp&rev=1.154&root=/cvsroot#2753
"// NS_STYLE_OVERFLOW_CLIP is a deprecated value, so if it's specified"
So that does mean one should not use NS_STYLE_OVERFLOW_CLIP anymore?

Comment 11

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

I'm not sure. It's slightly more efficient than overflow:hidden.

Comment 12

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

checked in

Comment 13

[Hideo Oshima]

I confirmed to solve the problem.
Thanks.

Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8b2) Gecko/20050412 Firefox/1.0+