44192 - scrollbars cover right/bottom of fixed positioned elements

Status: RESOLVED WORKSFORME

(Core :: Layout: Positioned, defect)

Description

[Jeff Bailey]

On Linux 2000062820, the right and bottom scrollbars cover text.  I have checked
this at 1024x768.

This is not a new problem.  I had thought originally thought this was part of
bug#36211, but the page for that now works for me correctly, so I'm now guessing
it's a different bug.

I'll will attach a simplified test case shortly.

Comment 1

[Jeff Bailey]

Attachment: This file works fine, and forms part of my comparison for expected behaviour

Comment 2

[Jeff Bailey]

Attachment: "columntest" is like "scrolltest" with a div added to shuffle all the text into column 2. The bug is triggered by this. I used column 2 to demonstrate both the side and bottom cut offs.

Comment 3

[David Krause]

confirming on 2000062808 linux

Comment 4

[Jeff Bailey]

Changing platform to all - I see this problem on Windows 98 2000062820 as well.
 Verifying this on Win98 is more difficult, because the text doesn't display at
first, until I grabbed the invisible scrollbar and moved it.  Then the scrollbar
and the text both appeared with the same symptoms as Linux.

Comment 5

[John Morrison]

 Thanks. We really have two problems here:

(1) the originally reported problem: I see in the second attachment, 
    that the gfx scrollbar indeed does cover the rightmost characters
    in the column. (i.e., the right edge of the wrapped text is located 
    underneath the scrollbar). [mac/linux/win32 -- win32 after you get
    the text to display].
 
    However, I'm not clear what you mean by 'bottom scrollbar covers
    text'; with 'overflow:scroll' the bottom scrollbar is created and 
    it must intersect with the overflowing text. However, when I scroll
    to the end of the text, none of that text is covered by the scrollbar
    (which is the correct result). 

(2) the followup problem that was noticed: on win32 (and on mac) there are 
    some painting problems. I will fill a separate bug for that. 

Note that with _native_ scrollbars on mac/win32/linux, the attached testcase
reveal a whole slew of different problems and the page is pretty much unusable 
(but the focus is on gfx scrollbars now, and native is slowly rotting). 

Passing on to evaughan (who may be slightly bummed to hear that there are a
couple of lingering problems).

[Tested with win95/mac/linux 2000062914 respin builds]

Comment 6

[Jeff Bailey]

Attachment: This is an image of what I see on my machine. The scroll bar is all the way to the bottom and I can highlight the text under the scroll bar. It has covered the words "fratrum nostrorum subscriptorum."

Comment 7

[John Morrison]

Ah, yes. I do see that the bottom scrollbar is covering the last line of 
text. (Hey. So how was I supposed to know that I was looking at an incomplete
Latin sentence :-). 

Comment 8

[Eric Vaughan]

fixed

Comment 9

[Jeff Bailey]

I still see this on Linux build 2000071120.  I will check again with tommorows
build on both Linux and Windows.

Comment 10

[Jeff Bailey]

Reopening bug - I still see this on Linux 2000071208, and Asa confirms 071208
NT4.  If it's not clear, (Bugzilla lays this out poorly on the page) the test
case is the second attachment 06 [details]/29/00 01:18.

Comment 11

[John Morrison]

moving evaughan non-nsbeta3+ bugs to Future milestone, per xptoolkit triage.

Comment 12

[Gervase Markham [:gerv]]

Note: this problem _not_ fixed by nsViewManager2 2001-03-24 Win95.

Gerv

Comment 13

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

*** Bug 89293 has been marked as a duplicate of this bug. ***

Comment 14

[Markus Gerstel]

Bug 89293 resolved itself somewhere between August 10th and 17th.

Why ? Nobody knows... My page works now as intended.

Comment 15

[Pierre Saslawsky]

*** Bug 46814 has been marked as a duplicate of this bug. ***

Comment 16

[Pierre Saslawsky]

There is a very nice and simple testcase at attachment 12075 [details] under bug 46814 
which was closed as dup.

Comment 17

[Boris Zbarsky [:bzbarsky]]

Attachment: minimal-ish testcase

Position:fixed is incidental here; I get the same results with
position:absolute for the outer div.  The key is the margin on the inner div --
we seem to not be taking that into account when computing the overflow area....

Comment 18

[Boris Zbarsky [:bzbarsky]]

Over to the right component.

Comment 19

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Do we have a bug somewhere on not including margins in overflow areas?  I think
so...

Comment 20

[Boris Zbarsky [:bzbarsky]]

There is bug 47710, but that's not quite the same... Note that we _do_ include
the top margin of an in-flow element in the overflow area.  The inner div being
absolutely positioned is necessary to cause the problem in this testcase.

I tried hacking the child bounds computations in nsAbsoluteContainingBlock by
moving over the child bounds by the size of the computed margin in Reflow(), but
that had no effect (it may be that Reflow() is not even being called here and
IncrementatlReflow() is what I should have poked at....)

Comment 21

[Steve England [:stevee]]

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8b2) Gecko/20050603
Firefox/1.0+ ID:2005060305
Is this bug still present? I am a little unclear what the testcase is meant to
show, but all 25 lines are present for me now.

Comment 22

[Boris Zbarsky [:bzbarsky]]

There's a good chance that roc's scrollframe changes helped here.

Comment 23

[Samuel Sidler (old account; do not CC)]

*** Bug 307071 has been marked as a duplicate of this bug. ***

Comment 24

[Gérard Talbot]

Isn't bug 211562 actually a duplicate of this bug? Just asking..

Comment 25

[Boris Zbarsky [:bzbarsky]]

I don't think it is -- one uses fixed pos, the other absolute pos.  They're using different parent frames, etc.

Comment 26

[Gérard Talbot]

Ok, Boris. Thank you for clarifying this.

Comment 27

[Gérard Talbot]

The URL (http://hurddocs.org/) is now "Server not found". 
So I try attachment 122417 [details] with Minefield/3.0a4pre
rv:1.9a4pre build Gecko/20070404 (under XP Pro SP2) and I can view/read the "Line 25": in other words, the last line of text is NOT covered by the bottom horizontal scrollbar. 
The other issue (5em bottom-margin of #column2) is bug 47710 as far as I understand this.
Is this bug fixed?

Comment 28

[Boris Zbarsky [:bzbarsky]]

Looks like it to me.