1214751 - [APZ] Cannot re-order tabs by drag & drop while elements of page are doing animation

Status: RESOLVED INCOMPLETE

(Core :: Panning and Zooming, defect)

Description

[Alice0775 White]

[Tracking Requested - why for this release]:


Steps to Reproduce:
1. Open 2 tabs [A] [B] 
2. Open https://www.mozilla.org/en-US/firefox/desktop/?utm_source=snippet&utm_medium=snippet&utm_campaign=default+feature+snippe in new tab[C]

3. Try drag tab[C] while the Firefox logo is animation.

4. optionally, reload the tab[C], and repeat Step3.

Actual Results:
Cannot re-order tabs by drag & drop

Comment 1

[Alice0775 White]

Application Basics
------------------

Name: Firefox
Version: 44.0a1
Build ID: 20151014030223
Update Channel: nightly
User Agent: Mozilla/5.0 (X11; Linux i686; rv:44.0) Gecko/20100101 Firefox/44.0
Multiprocess Windows: 1/1 (default: true)
Safe Mode: false

Crash Reports for the Last 3 Days
---------------------------------

Report ID: bp-ef9cf9a7-1d11-4a0e-87ce-474022151013
Submitted: 1 day ago

Report ID: bp-0e700bcb-dd9e-4a3a-8d8b-2ef9e2151011
Submitted: 3 days ago

All Crash Reports

Extensions
----------

Name: Ubuntu Modifications
Version: 3.2
Enabled: false
ID: [e-mail]

Graphics
--------

Adapter Description: VMware, Inc. -- Gallium 0.4 on SVGA3D; build: RELEASE;
Asynchronous Pan/Zoom: wheel input enabled
Device ID: Gallium 0.4 on SVGA3D; build: RELEASE;
Driver Version: 2.1 Mesa 10.1.3
GPU Accelerated Windows: 0/1 Basic (OMTC)
Supports Hardware H264 Decoding: No;
Vendor ID: VMware, Inc.
WebGL Renderer: VMware, Inc. -- Gallium 0.4 on SVGA3D; build: RELEASE;
windowLayerManagerRemote: true
AzureCanvasBackend: cairo
AzureContentBackend: cairo
AzureFallbackCanvasBackend: none
AzureSkiaAccelerated: 0
CairoUseXRender: 1

Important Modified Preferences
------------------------------

browser.cache.disk.capacity: 1048576
browser.cache.disk.filesystem_reported: 1
browser.cache.disk.smart_size.first_run: false
browser.cache.frecency_experiment: 2
browser.download.importedFromSqlite: true
browser.places.smartBookmarksVersion: 7
browser.startup.homepage_override.buildID: 20151014030223
browser.startup.homepage_override.mstone: 44.0a1
dom.apps.reset-permissions: true
dom.mozApps.used: true
extensions.lastAppVersion: 44.0a1
network.cookie.prefsMigrated: true
places.history.expiration.transient_current_max_pages: 33583
plugin.disable_full_page_plugin_for_types: application/pdf
privacy.sanitize.migrateClearSavedPwdsOnExit: true

Important Locked Preferences
----------------------------

JavaScript
----------

Incremental GC: true

Accessibility
-------------

Activated: false
Prevent Accessibility: 0

Library Versions
----------------

NSPR
Expected minimum version: 4.10.9
Version in use: 4.10.9

NSS
Expected minimum version: 3.20 Basic ECC
Version in use: 3.20 Basic ECC

NSSSMIME
Expected minimum version: 3.20 Basic ECC
Version in use: 3.20 Basic ECC

NSSSSL
Expected minimum version: 3.20 Basic ECC
Version in use: 3.20 Basic ECC

NSSUTIL
Expected minimum version: 3.20
Version in use: 3.20

Experimental Features
---------------------

Sandbox
-------

Seccomp-BPF (System Call Filtering): true
Seccomp Thread Synchronization: true
User Namespaces: true
Media Plugin Sandboxing: true

Comment 2

[Alice0775 White]

Repeat Stap3-4, Finally , D&D tab reorder is fully broken.

Comment 3

[Alice0775 White]

Regression window:
https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=ccd3484ebb4c742a43b5d938911eac2d7e670d43&tochange=97e537f85183


Regressed by:
30742281c223	Kartikaya Gupta — Bug 1143856 - Enable APZ on Linux desktop nightly builds. r=botond

Setting layers.async-pan-zoom.enabled = false fixes the problem.

Comment 4

[:Gijs (out for now; he/him)]

(In reply to Alice0775 White from comment #3)
> Regression window:
> https://hg.mozilla.org/mozilla-central/
> pushloghtml?fromchange=ccd3484ebb4c742a43b5d938911eac2d7e670d43&tochange=97e5
> 37f85183
> 
> 
> Regressed by:
> 30742281c223	Kartikaya Gupta — Bug 1143856 - Enable APZ on Linux desktop
> nightly builds. r=botond
> 
> Setting layers.async-pan-zoom.enabled = false fixes the problem.

kats, can you look into what's going on here?

Comment 5

[Jan Alexander Steffens [:heftig]]

I remember some recent trouble dragging tabs. Does disabling nglayout.enable_drag_images help?

Comment 6

[Alice0775 White]

(In reply to Jan Steffens from comment #5)
> I remember some recent trouble dragging tabs. Does disabling
> nglayout.enable_drag_images help?

The pref does not help.

Comment 7

[Botond Ballo [:botond]]

I tried reproducing this, but wasn't able to. The rendering of the tab bar and of the thumbnail of the page is a bit slow while dragging (but then I'm testing a debug build), but otherwise I'm able to re-order tabs by dragging fine.

Comment 8

[Chris AtLee [:catlee]]

I cannot drag-and-drop tags in 2015-10-19's build on Linux. Sometimes a small preview window pops up, sometimes not.

Comment 9

[Alice0775 White]

Attachment: log.txt

log with NSPR_LOG_MODULES=nsDragServic

Attempt to drag tab and repeat.  Finally, I cannot drag tab anymore.

Comment 10

[Alice0775 White]

Attachment: log with NSPR_LOG_MODULES=nsDragService:5

log with NSPR_LOG_MODULES=nsDragService:5 of shell env.

Attempt to drag tab and repeat.  and then finally, completely it was no longer able to drag tabs.

Comment 11

[Karl Tomlinson (:karlt)]

The patch at https://hg.mozilla.org/try/rev/659a8bf97a13 was for bug 1216916, which exists sometimes without APZ.
The log here is consistent with a similar state, but I don't know what triggers that state.

Comment 12

[Kartikaya Gupta (email:[e-mail])]

Thanks for the log, Alice! I'm not able to reproduce this locally but I'll take a look at the log and see what I can come up with. I might need more logs later once I narrow it down. Assigning to myself for now.

Comment 13

[Kartikaya Gupta (email:[e-mail])]

Those of you who can reproduce this, are you on a HiDPI display?

Comment 14

[Ritu Kothari (:ritu) (Inactive, please n-i to RyanVM, jcristau, or pascal)]

Tracked for 44 for now but if we cannot get a consistent repro/root cause on this one, I might decide to untrack it eventually.

Comment 15

[Kartikaya Gupta (email:[e-mail])]

Note that if indeed it is a result of APZ then it should be nightly-only and so shouldn't affect 44 any more.

Comment 16

[Chris AtLee [:catlee]]

This is still impacting me on Nightly (45).

My DPI is 96x96.

Comment 17

[Kartikaya Gupta (email:[e-mail])]

:catlee showed me this just now, and we confirmed it still happens with APZ disabled, at least on his machine. This contradicts what Alice found in comment 3, so maybe there are multiple issues here. :catlee will try to get a regression window for his issue, maybe it will shed more light on the matter.

Comment 18

[Chris AtLee [:catlee]]

Using mozregression, I found this regression range:

https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=7042219cf7f332ce97605958506557a9ae7e9402&tochange=884385e6f556fc07d33ddd790474d459a7e55aab

Comment 19

[Kartikaya Gupta (email:[e-mail])]

Gabor, do you know why bug 863512 might have caused this? ^

Comment 20

[Alice0775 White]

(In reply to Chris AtLee [:catlee] from comment #18)
> Using mozregression, I found this regression range:
> 
> https://hg.mozilla.org/integration/mozilla-inbound/
> pushloghtml?fromchange=7042219cf7f332ce97605958506557a9ae7e9402&tochange=8843
> 85e6f556fc07d33ddd790474d459a7e55aab

This is different bug, see Bug 1212733.

Comment 21

[Kartikaya Gupta (email:[e-mail])]

Ah, thanks! Looks like a fix for that just landed on inbound. It'll be interesting to see if that resolves this bug (i.e. comment 0) as well.

Comment 22

[Chris AtLee [:catlee]]

This is working in today's nightly. The small previews of the tabs are sometimes blank, but I can rearrange them now.

Comment 23

[Kartikaya Gupta (email:[e-mail])]

Thanks, sounds like your issue was bug 1212733.

Alice, can you check if the issue from comment 0 still happens?

Comment 24

[Alice0775 White]

(In reply to Kartikaya Gupta (email:[e-mail]) from comment #23)
> Thanks, sounds like your issue was bug 1212733.
> 
> Alice, can you check if the issue from comment 0 still happens?

Hmm, I can still reproduce the problem.
I cannot re-order tabs by drag & drop while elements of page are doing animation .

https://hg.mozilla.org/mozilla-central/rev/eb3016abd37db2e6a6d923265047e84b12c0af61
Mozilla/5.0 (X11; Linux i686; rv:45.0) Gecko/20100101 Firefox/45.0 ID:20151118030232

Comment 25

[Alice0775 White]

The build seems to be fixed problem of Comment#2.

Is the following problem by design? 
"Cannot re-order tabs by drag & drop while elements of page are doing animation ."

Comment 26

[Kartikaya Gupta (email:[e-mail])]

I don't think it's by design, it's probably a bug. But if it only happens during animation it's less severe, so it seems like bug 1212733 fixed the largest part of the problem.

Comment 27

[Alice0775 White]

OK.

Comment 28

[Milan Sreckovic [:milan] (needinfo for best results)]

Sounds like we won't fix for 44, we can stop tracking because the issue is less severe than before?

Comment 29

[Kartikaya Gupta (email:[e-mail])]

It should only affect 46+ now, since that's where APZ is enabled. And yes, it is less severe than originally reported so it's worth re-evaluating the tracking status. I'm shifting the flag from tracking-44+ to tracking-46?

Comment 30

[Liz Henry (:lizzard) (relman/hg->git project)]

Tracking for 46. mstange can you help find an owner for this regression?

Comment 31

[Markus Stange [:mstange]]

Forwarding to kats - not sure what we can do here as long as we can't reliably reproduce it.

The only idea I have is that dragging might require the main thread to be responsive, and that the animation might be painting more because of the larger displayport, and that this puts too much strain on the main thread... but that's not really actionable. Or we might be painting through xrender, and since we're painting more, this might put too much strain on the window server or something. We should check whether turning off xrender (which nical is trying out in some other bug at the moment) helps with this.

Comment 32

[Kartikaya Gupta (email:[e-mail])]

I tried again to reproduce this issue, and was not able to. Looking through Alice's log there are 30+ drag sessions recorded, and it's not clear to me which one of those is the relevant drag session so I can't really use that information either. Normally I would expect the last drag session to be the one that needs debugging but in this case that looks like most of the previous ones so I don't think that's it. Alice, is it possible for you to get another log with NSPR_LOG_MODULES=nsDragService:5, but identify the log section for a single failed drag?

One way to do this would be to have the log go to stdout or stderr, and copy the program output upto the point that the bug has been reproduced, without copying anything from after that point. That way the relevant log output will be at the very end of the file, and I'll still have the history of all the output from the start of the browsing session. There might be other assertions/warnings that might be useful in that. Thanks.

Comment 33

[Alice0775 White]

(In reply to Kartikaya Gupta (email:[e-mail]) from comment #32)
> I tried again to reproduce this issue, and was not able to. Looking through
> Alice's log there are 30+ drag sessions recorded, and it's not clear to me
> which one of those is the relevant drag session so I can't really use that
> information either. Normally I would expect the last drag session to be the
> one that needs debugging but in this case that looks like most of the
> previous ones so I don't think that's it. Alice, is it possible for you to
> get another log with NSPR_LOG_MODULES=nsDragService:5, but identify the log
> section for a single failed drag?
> 
> One way to do this would be to have the log go to stdout or stderr, and copy
> the program output upto the point that the bug has been reproduced, without
> copying anything from after that point. That way the relevant log output
> will be at the very end of the file, and I'll still have the history of all
> the output from the start of the browsing session. There might be other
> assertions/warnings that might be useful in that. Thanks.

last one. 

I can still repro the problem. Maybe the reason is poor CPU and blocked GPU.
And there is no duplicate bug.
So, I think this is no need track.

Kartikaya,
sorry, I close this as incomplete.