803170 - Settings sub-panels no longer load

Status: VERIFIED FIXED

(Firefox OS Graveyard :: Gaia, defect, P1)

Description

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

STR
 (1) Open settings app
 (2) Tap "Wifi"

See white panel.

This is borderline blocker but devices that are already configured will continue to work.

inbound 06777f847494, gaia 3bdf9c46738a4bab9485552fc28708ac1603a73f

Comment 1

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

This is a gecko regression.

Comment 2

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

Regression range

http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=0923556aea0e&tochange=06777f847494

Comment 3

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

http://hg.mozilla.org/integration/mozilla-inbound/rev/fbf63fe2c1b0

Pretty scary regression, looking ...

Comment 4

[Naoki Hirata :nhirata (please use needinfo instead of cc)]

Hrm. odd.  I didn't see it on today's nightly build for otoro:

Otoro phone, build 2012-10-18
Taken from default.xml in b2g-distro: 
* "platform_build" revision= 0d6d050bc37d3167cc82a1885fd7660456bb0f4e
* "gaia" revision= e3efbd0411218762cf9a62278bf58ee513ff331f 
* "releases-mozilla-aurora" revision= 38c06c8b4f8f8a6a513a66ab62c9421835017c9f
* "gonk-misc" revision= 63287df891a154aff6f3c58587e739fa3f1684eb

gaia didn't seem to change.  I twas only the aurora and gonk from yesterday.

Comment 5

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

We end up computing a transform for a content layer that makes it invisible

696526656[126f850]:           OGLShadowContainerLayer (0x2e6e928) [shadow-transform=[ 1 0; 0 1; 0 20; ]] [shadow-visible=< (x=0, y=0, w=320, h=460); >] [visible=< (x=0, y=0, w=320, h=460); >] [metrics={ viewport=(x=0.000000, y=0.000000, w=320.000000, h=460.000000) viewportScroll=(x=320.000000, y=0.000000) displayport=(x=0.000000, y=0.000000, w=0.000000, h=0.000000) scrollId=1 }]
696526656[126f850]:             OGLShadowThebesLayer (0x2a0be88) [shadow-transform=[ 1 0; 0 1; -320 0; ]] [transform=[ 1 0; 0 1; -320 0; ]] [not visible] [opaqueContent]

Will try to make a testcase and poke a bit, but I think the risk to web compat of this is pretty high so might need backing out if a fix isn't conjured quickly.

Comment 6

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

s/a content layer/the content layers/

Comment 7

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

I was unable to get the settings app working standalone.  Grrr.

Comment 8

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

(In reply to Naoki Hirata :nhirata from comment #4)
> Hrm. odd.  I didn't see it on today's nightly build for otoro:
> 
> * "releases-mozilla-aurora" revision=
> 38c06c8b4f8f8a6a513a66ab62c9421835017c9f

Right, this commit isn't on aurora.

Comment 10

[Naoki Hirata :nhirata (please use needinfo instead of cc)]

oh right.  dev are working on nightly builds... hrm.  that complicates things because QA tests on aurora...

Comment 11

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

QA should absolutely stay on aurora.  Developers should focus on aurora too, but we need to dual-land so need to be able to test on inbound/m-c.  This bug makes that very hard / impossible.

Comment 12

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

Attachment: Small test

Exhibits the same symptom as settings app.

Comment 13

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

So before bug 703241, this test works well, no janking with empty space.

In gecko after bug 703241, the first transition leaves us in white space with no way to go back anywhere.

In Chrome 23, the transition from the #root anchor to #sub is through empty space, but when the transition finishes, the #sub content shows up again.  So still kind of broken but less so.

kaze, I think we're going to want to change how this transition works since it appears to be slightly fragile.

roc, what do you want to do about the behavioral discrepancy?

Comment 14

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

Hm, although I have a rather evil potential workaround ...

Comment 15

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

  window.onhashchange = function() { window.scrollTo(0, 0); }

ftw

Comment 16

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

(FTR, that also makes the animation not jank in Chrome, too.)

Comment 17

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

Attachment: Band-aid to fix transitions to sub-panels in settings app

https://github.com/mozilla-b2g/gaia/pull/5890

Comment 18

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

Followup bug 803409.

Comment 19

[Fabien Cazenave [:kaze]]

(In reply to Chris Jones [:cjones] [:warhammer] from comment #13)
> kaze, I think we're going to want to change how this transition works since
> it appears to be slightly fragile.

Agreed.

Comment 20

[Deepa Subramanian]

verified on unagi build id:  20130103070201