265860 - Links "jump" out from under mouse onclick

Status: RESOLVED FIXED

(Core :: Layout, defect)

Description

[Jerry Baker]

There are some pages where links "jump" out from under the mouse on the first
click. No JS and no CSS is set on these links.

One example:
http://www.xs4all.nl/~eijsel/inhoud.html

Click on any link *in the first paragraph* (e.g., "Birdphotography tips,"
"Hotspots," "Photo Database."

Comment 1

[Martijn Wargers (dead)]

I can see the bug with:
Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8a5) Gecko/20041024
Firefox/0.9.1+
The <img> in the page has align="left". There is a chance this gets fixed with
the fix for bug 209694.

Comment 2

[Peter van der Woude [:Peter6]]

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.7.3) Gecko/20041024 Firefox/1.0

I'm not seeing anything unusual when I click a link

Comment 3

[Martijn Wargers (dead)]

It only happens in trunk (1.8a) builds.

Comment 4

[James Norris]

(In reply to comment #3)
> It only happens in trunk (1.8a) builds.

I have this problem, I believe, on Linux. The mouse cursor jumps about an inch
to the right on clicking a link. Or is this a seperate bug that I should file?

Comment 5

[OstGote!]

WFM with Mozilla 1.8a6 build 2004112806 on WinNT4.

Comment 6

[Martijn Wargers (dead)]

Attachment: Beginning of a testcase

Well, I can still see the "linkjumping" bug with a 2004-11-28 trunk build.
You can see it with this testcase when clicking on the "my pictures" link.
This testcase might very well be font/resolution dependent (so not a really
ideal testcase). My resolution is 1024x768.

Comment 7

[Troodon]

It's funny, but I discovered this bug when focusing a link listed on bug 261196
(allways reproducible).

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8a6) Gecko/20041203

Comment 8

[Robert Kaiser]

This is not a Firefox bug, as it seems, comments here actually never talk of FF
at all but only Mozilla suite builds.
Reassigning to a fitting core/layout component.

I'm seeing lots of jumping of boxes on http://www.fynf.at/ (just try to click on
any link in the middle area of that site and see wild jumps of the content), I
never managed to do a reduced testcase starting from there though.

Comment 9

[Jerry Baker]

(In reply to comment #8)
> This is not a Firefox bug, as it seems, comments here actually never talk of FF
> at all but only Mozilla suite builds.
> Reassigning to a fitting core/layout component.

This bug was reported for and from Firefox. Also, see comment #1.

Comment 10

[Martijn Wargers (dead)]

No, I tested it with Firefox, but this is typically a layout issue, and thus
this problem also happens with Mozilla. See also the other bugs that are
blocking bug 261196, all layout bugs.

Comment 11

[Jerry Baker]

I just want to be clear. When there is a bug filed against Mozilla, we are
supposed to file a duplicate bug against Firefox (because it can be fixed in one
and not the other), but if a bug is filed against Firefox it automatically
counts against Mozilla also?

Comment 12

[Robert Kaiser]

(In reply to comment #11)
> I just want to be clear. When there is a bug filed against Mozilla, we are
> supposed to file a duplicate bug against Firefox (because it can be fixed in one
> and not the other), but if a bug is filed against Firefox it automatically
> counts against Mozilla also?

For UI bugs (and all things that are built from different sources in FF and the
suite), it's two different bugs and two bugs should be filed.

For core things like Gecko/Layout, it's one source and one bug, and that's how
it should be filed. Now that we have the "Core" component in Bugzilla, you can
roughly suppose all bugs in that product apply to both products, while the bugs
in the othere seperate products apply only to those products.

For my fynf.at jumping, I believe it's a different bug as I can't see the
jumping in the original case here, and I can't see mine with 1.7.x but with
1.8a, so I'll file a new one...

Comment 13

[Vincent Lefevre]

Same problem (using the testcase) under Linux/PPC, using Mozilla (CVS checkout
and compiled a few hours ago). Hardware/OS should be set to All/All.

Comment 14

[Mats Palmgren (inactive)]

I can reproduce the bug up to and including 2005-01-27-10 trunk Linux.
Starting with 2005-01-28-05 and forward the bug does not occur.
In that span the only bugs that seems likely are bug 279786 or bug 273946.
Maybe bug 279786 fixed or wallpapered the problem?

Comment 15

[Boris Zbarsky [:bzbarsky]]

doesn't seem that likely....  I can see bug 273946 wallpapering it, though.

Comment 16

[Ronald Tilby]

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8b2) Gecko/20050409
Testcase WFM using Mozilla Trunk Nightly 2005040905 on Windows XP.

Comment 17

[Martijn Wargers (dead)]

Yes, this seems to be fixed. I can't reproduce with the 2005-04-08 build, not
with the testcase nor with the url.
Feel free to reopen if someone can still reproduce with the latest nightly build.

Comment 18

[Jerry Baker]

Happens on this page: http://www.newscientist.com/channel/space/mg18624945.400
using the 20050408 trunk. Click on the link in the story text.

Comment 19

[Jerry Baker]

Another site is http://www.data.scec.org/recenteqs/. Click on the map to zoom
in. When you click on the zoomed-in map, it jumps out from under your mouse.

Comment 20

[Martijn Wargers (dead)]

(In reply to comment #19)
> Another site is http://www.data.scec.org/recenteqs/. Click on the map to zoom
> in. When you click on the zoomed-in map, it jumps out from under your mouse.
This happpens because the Mozilla tries to get the whole link in sight when
focusing. That's a different bug (don't know the bug number).
Jerry, can you still see the bug with a build which has the fix for bug 240276
built in?

Comment 21

[Steve England [:stevee]]

Testcase WFM Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.7.7)
Gecko/20050414 Firefox/1.0.3
Testcase WFM Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8b2)
Gecko/20050429 Firefox/1.0+

Comment 22

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

Please file new bugs for any issues not yet resolved.