614467 - Tabs scattered all over the place when entering Panorama

Status: VERIFIED FIXED

(Firefox Graveyard :: Panorama, defect, P2)

Description

[(no longer active)]

When I tried to enter Panorama today, this is what I saw: http://grab.by/7yr2

Exiting and reentering Panorama just brings me to the exact same screen.  I didn't see anything on the error console.

Is this known?

Comment 1

[The 8472]

Did you restore a session? And is this on B7 or a current nightly?

Comment 2

[(no longer active)]

This is a current nightly, built from http://hg.mozilla.org/mozilla-central/rev/4d73807bfec7.

I always restore my session, but I last restarted my browser late last night or early this morning, IIRC, and have been heavily using it.

I tried grabbing the resizers at the bottom edge of the Mozilla, Travel, and the unnamed group (the big one at the center) to make them a few (maybe around 10) pixels bigger in each dimension, and right when resizing all of the tabs belonging to each group fell into their proper spot.  This is what I'm seeing now:

http://grab.by/7yrv

So I think the code responsible for figuring out whether each group is big enough for its tabs had somehow choked in the middle of its thing.

I wasn't paying close attention when I first entered Panorama, but I _think_ that the tabs were animating in seemingly random directions and they suddenly stopped.  But take this comment with a grain of salt (you'll be relying on my memory after all!!!)

Comment 3

[Michael Yoshitaka Erlewine [:mitcho]]

(In reply to comment #0)
> When I tried to enter Panorama today, this is what I saw: http://grab.by/7yr2
> 
> Exiting and reentering Panorama just brings me to the exact same screen.  I
> didn't see anything on the error console.
> 
> Is this known?

Off the top of my head, this looks more like an issue with restoring and placing the tabs while not yet having proper group information. The fact that you could jiggle the group bounds, forcing an .arrange(), and it worked, means that the shouldStack() and group info ultimately are working... it just seems like maybe it didn't have access to the group info right when you first launched panorama.

Did you launch Panorama right after startup?

Comment 4

[(no longer active)]

(In reply to comment #3)
> (In reply to comment #0)
> > When I tried to enter Panorama today, this is what I saw: http://grab.by/7yr2
> > 
> > Exiting and reentering Panorama just brings me to the exact same screen.  I
> > didn't see anything on the error console.
> > 
> > Is this known?
> 
> Off the top of my head, this looks more like an issue with restoring and
> placing the tabs while not yet having proper group information. The fact that
> you could jiggle the group bounds, forcing an .arrange(), and it worked, means
> that the shouldStack() and group info ultimately are working... it just seems
> like maybe it didn't have access to the group info right when you first
> launched panorama.

That seems plausible, indeed.

> Did you launch Panorama right after startup?

No, quite a while after startup in fact.

Comment 5

[matthias koplenig [:metasieben]]

wondering if this could be somehow related to setting >browser.sessionstore.max_concurrent_tabs< to 0, as all the screenshots in this bug, as well as my own[1] don't show any tab-previews.


[1] http://grab.by/7D9S

Comment 6

[timeless]

Attachment: my overflow (panorama started long after session started which was a restore)

Comment 7

[Kevin Hanes]

bugspam (moving b9 to b10)

Comment 8

[Kevin Hanes]

bugspam (removing b9)

Comment 9

[Michael Yoshitaka Erlewine [:mitcho]]

Does this still happen? We've landed some better session store integration recently... I wonder if this might be fixed.

Comment 10

[matthias koplenig [:metasieben]]

can confirm that the _scattering_ is fixed.

however this http://grab.by/8k9A does still happen, wondering if this is related?

Comment 11

[Emanuele Alimonda]

The scattering seems to always happen in case, before you enter Panorama and after restoring a saved session, you resize the Firefox window smaller than it was when you saved the session.  I'm still getting it on the 2011-01-11 nightly (only tested on Mac OS X 10.6).

When you enter Panorama, groups get resized to fit the current window size, while tab thumbnails are kept in their original position.  Once switching to a group by clicking on one of the thumbnails, all the tabs for that group get redrawn inside the group bounds.  Tabs get also redrawn in the correct position by resizing the Firefox window when in Panorama.

I originally thought it was due to incompatibilities with the BarTab extension, but a restart without BarTab proved me wrong, as it still happened.  In my case, browser.sessionstore.max_concurrent_tabs is set to the default value (3).

http://img89.imageshack.us/img89/1667/screenshot20110111at102.png

I also occasionally get something similar to what shown in Comment #10.

By the way, why was Platform set to x86 Windows XP?  All the screenshots I see on this bug appear to have been taken on Mac OS.

Comment 12

[Michael Yoshitaka Erlewine [:mitcho]]

(In reply to comment #10)
> can confirm that the _scattering_ is fixed.
> 
> however this http://grab.by/8k9A does still happen, wondering if this is
> related?

Thanks Matthias. This particular thing, I believe, will be fixed with bug 622835.

Comment 13

[Michael Yoshitaka Erlewine [:mitcho]]

(In reply to comment #11)
> The scattering seems to always happen in case, before you enter Panorama and
> after restoring a saved session, you resize the Firefox window smaller than it
> was when you saved the session.

Ah. If this is the only way to reproduce this, though, it may be fixed by bug 601534.

> By the way, why was Platform set to x86 Windows XP?  All the screenshots I see
> on this bug appear to have been taken on Mac OS.

Thanks. Fixed.

Comment 14

[Ian Gilman [:iangilman]]

(In reply to comment #13)
> (In reply to comment #11)
> > The scattering seems to always happen in case, before you enter Panorama and
> > after restoring a saved session, you resize the Firefox window smaller than it
> > was when you saved the session.
> 
> Ah. If this is the only way to reproduce this, though, it may be fixed by bug
> 601534.

We shouldn't wait for that bug though.

David, are you working on this, or should someone else take it?

Comment 15

[David Dahl :ddahl]

Someone can take it.

Comment 17

[Michael Yoshitaka Erlewine [:mitcho]]

(In reply to comment #14)
> > Ah. If this is the only way to reproduce this, though, it may be fixed by bug
> > 601534.
> 
> We shouldn't wait for that bug though.

Indeed. We just punted on 601534.

Comment 18

[Tim Taubert [:ttaubert] (inactive)]

This is fixed when patch for bug 624953 lands.

Comment 19

[Michael Yoshitaka Erlewine [:mitcho]]

softblocker = critical

Comment 20

[Tracy Walker [:tracy]]

verified with nightly build of minefield.