John Reporter
	Description • 17 years ago

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9b5pre) Gecko/2008032304 Minefield/3.0b5pre
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9b5pre) Gecko/2008032304 Minefield/3.0b5pre

If you look certain websites, you will notice that there is a one pixel gap after the background image fades into a solid color background.

Example: http://portfolioso.com

The example site is valid XHTML 1.0 Strict and valid CSS 2.1 and this did not happen in yesterday's nightly or on other browsers. Someone with Vista said it also did not happen.

Regression:
Works: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9b5pre) Gecko/2008032204 Minefield/3.0b5pre
Broken: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9b5pre) Gecko/2008032304 Minefield/3.0b5pre 

Reproducible: Always

Steps to Reproduce:
1. Visit http://portfolioso.com
2. Notice the rendering issue
Actual Results:  
Page rendering: A gap appears where it should not

Expected Results:  
There should not be a gap after a background image

Screenshot: http://img139.imageshack.us/my.php?image=issueme5.png

	Louis
	Comment 1 • 17 years ago

Works fine for me. I do not see the pixel gap.

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9b5pre) Gecko/2008032304 Minefield/3.0b5pre ID:2008032304

Adblock Plus 0.7.5.3
Brief 1.1
CustomizeGoogle 0.71
Fission 0.9.5
IE Tab 1.5.20080310
Metrics Collector 2a6
Nightly Tester Tools 1.3
OpenDownload 1.0.0
ShowIP 0.8.05

	John Reporter
	Comment 2 • 17 years ago

Seems to be a corrupted profile. Odd

	John Reporter
	Comment 3 • 17 years ago

Edit: This happens when you zoom out one level. It remembered my page zoom settings and that was doing it. Therefore, it still is a bug.

	John Reporter
	Comment 4 • 17 years ago

(In reply to comment #2)
> Seems to be a corrupted profile. Odd

Ignore that

(In reply to comment #3)
> Edit: This happens when you zoom out one level. It remembered my page zoom
> settings and that was doing it. Therefore, it still is a bug.
> 

This the regression range I gave is wrong. Please disregard.

	John Reporter
	Comment 5 • 17 years ago

Confirmed Regression range (it's from a while ago):

WORKS: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a9pre) Gecko/2007102504 Minefield/3.0a9pre

BROKEN: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a9pre) Gecko/2007102605 Minefield/3.0a9pre

	John Reporter
	Comment 6 • 17 years ago

One other thing. I never actually made the steps to reproduce clear since I didn't understand the bug at first. But now I narrowed it down:

Steps to Reproduce:
1. Visit http://portfolioso.com
2. Zoom out once (control -)
Actual Results:  
Page rendering: A gap appears under the menu (where the background image ends). There should not be a gap here on page zoom.

	Peter van der Woude [:Peter6]
	Comment 7 • 17 years ago

range from comment 6
http://bonsai.mozilla.org/cvsquery.cgi?treeid=default&module=PhoenixTinderbox&branch=HEAD&branchtype=match&filetype=match&whotype=match&sortby=Date&hours=2&date=explicit&mindate=2007-10-25+04%3A00%3A00&maxdate=2007-10-26+04%3A00%3A00&cvsroot=%2Fcvsroot

	-fullmetaljacket-
	Comment 8 • 17 years ago

bug 389628?

	Mats Palmgren (inactive)
	Comment 10 • 17 years ago

Also occurs on Linux and MacOSX.

	Mats Palmgren (inactive)
	Comment 11 • 17 years ago

Since we're rounding the pattern origin (p0), seems like we should adjust
the size accordingly?  It fixes the problem on Linux and OSX, haven't
tested Windows.  (passes reftests on Linux, fwiw)

	Robert O'Callahan (:roc) (email my personal email if necessary)
	Comment 12 • 17 years ago

I don't seem to see this bug on Mac. I'm not sure why the offset is rounded at all, to be honest.

	John Reporter
	Comment 13 • 17 years ago

(In reply to comment #12)
> I don't seem to see this bug on Mac. I'm not sure why the offset is rounded at
> all, to be honest.
> 

Zooming out on the testcase breaks it. It seems to work if you zoom in. Also, is there any way I can get a compiled version of this patch so I can test it on Windows and let you know if it should be checked in?

	John Reporter
	Comment 14 • 17 years ago

Can someone commit the fix or come up with a test build, so I can test this on Windows? Thanks

	Robert O'Callahan (:roc) (email my personal email if necessary)
	Comment 15 • 16 years ago

The patch in bug 430465 might fix this.

However, we aren't going to block on it.

	Mats Palmgren (inactive)
	Comment 16 • 16 years ago

It didn't fix it (at least not on Linux).

	Mats Palmgren (inactive)
	Comment 18 • 16 years ago

(In reply to comment #14)
Test builds are available here:
https://build.mozilla.org/tryserver-builds/2008-04-25_03:30-mpalmgren@mozilla.com-424672_patch2/

	John Reporter
	Comment 19 • 16 years ago

Fixed in the test build posted above.

	Justin Dolske [:Dolske]
	Comment 20 • 16 years ago

checkin-needed is only for reviewed patches.

	Dão Gottwald [:dao]
	Comment 21 • 16 years ago

Is this ready for review?

	Mats Palmgren (inactive)
	Comment 22 • 16 years ago

Updated to tip.

	Mats Palmgren (inactive)
	Comment 23 • 16 years ago

(In reply to comment #21)
> Is this ready for review?

I wanted to see the outcome of bug 430465 first, but perhaps we
should take this in the meantime?  It passes reftests on Linux,
Windows and MacOSX.

	Robert O'Callahan (:roc) (email my personal email if necessary)
	Comment 24 • 16 years ago

What's the rationale for this patch? I think it should be solved at a higher level, like bug 430465 or something similar.

	Mats Palmgren (inactive)
	Comment 25 • 16 years ago

(In reply to comment #24)
> What's the rationale for this patch?

It's just an intermediate fix until we have something better.
Bug 430465 didn't seem to fix it when I tested it (comment 16).

	Robert O'Callahan (:roc) (email my personal email if necessary)
	Comment 26 • 16 years ago

I'd prefer not to take an intermediate fix unless this is super-critical.

	Mats Palmgren (inactive)
	Comment 27 • 16 years ago

Comment on attachment 336871 [details] [diff] [review]
Possible fix?

Ok, it's not critical at all.

	John Reporter
	Comment 28 • 16 years ago

Whatever was checked in recently seemed to have fixed this. Marking resolved wontfix, since another fix took care of the problem.

	:Gavin Sharp [email: gavin@gavinsharp.com]
	Comment 29 • 16 years ago

WONTFIX means "we won't fix it", WORKSFORME is the appropriate resolution for this case.

	ram
	Comment 30 • 16 years ago

hfhs