992263 - regression found on datazilla in cold_load_time test for most apps

Status: RESOLVED DUPLICATE of bug 987994

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

Description

[Joel Maher ( :jmaher ) (UTC -8)]

here is a link to datazilla:
https://datazilla.mozilla.org/b2g/?branch=master&device=hamachi&range=30&test=cold_load_time&app_list=contacts&app=contacts&gaia_rev=eee8caa81a368f0f&gecko_rev=df7b26e90378&plot=median

the regression is found in:
https://github.com/mozilla-b2g/gaia/commit/8248f8da7a29757d

and gecko revision:
http://git.mozilla.org/?p=releases/gecko.git;a=commit;h=df7b26e90378

Comment 1

[Ben Kelly [:bkelly, not reviewing]]

Thanks for filing this Joel!

Comment 2

[Mason Chang [Inactive] [:mchang]]

Triage: QA, please find a regression window if possible. Thanks!

Comment 3

[Jason Smith [:jsmith]]

(In reply to Mason Chang [:mchang] from comment #2)
> Triage: QA, please find a regression window if possible. Thanks!

That's not really in our scope to do - perf regression windows are primarily driven by the performance team, not QA.

Comment 4

[bhavana bajaj [:bajaj]]

(In reply to Mason Chang [:mchang] from comment #2)
> Triage: QA, please find a regression window if possible. Thanks!

:mchang can you help how heavily we've regressed the numbers here to actually block on this bug for 2.0 ? sry, may need more practice reading datazilla here if its quite obvious, so a bit hand holding may help here until I get there :) Also given we have this bug filed and commits mentioned in the description, should we try backing those out to see if it helps here ?

Comment 5

[Mason Chang [Inactive] [:mchang]]

It looks like it is a 700ms regression. You can check this by seeing the end line, read the average (1559) from the starting point (750ms) and subtract the difference. Need info huseby / jhylands to see if someone can investigate.

Comment 6

[Jon Hylands [:jhylands]]

I've bisected this, and came up with this commit:

439e40ef0e79393f94ca102b9fbd0785355e4b34 is the first bad commit
commit 439e40ef0e79393f94ca102b9fbd0785355e4b34
Author: Alive Kuo <[e-mail]>
Date:   Mon Feb 3 11:49:44 2014 +0100

    Bug 961800 - Implement Child Window Factory.

    Conflicts:
        apps/system/js/app_transition_controller.js
        apps/system/js/app_window.js
        apps/system/js/app_window_manager.js
        apps/system/js/cards_view.js
        apps/system/style/window.css

:100644 100644 635b6e3f784982c7c3944ba50a05a3259b8a8fab 11c79904ce42b229c18a28d24c9bd20ed718dda8 M      Makefile
:040000 040000 e3b5cbb0db28d2a6e75962207f44dc2a630d7e10 d3cc2b91fe4a8ae471ab9e2f3ee3eb7c0f246305 M      apps
:040000 040000 c787ffbabb879c3fc111407aefdefa257f68b019 b27d5e19622014e8f61747a18627aa521d5a890b M      build
:040000 040000 4ca18cee8ccb928feb6bd18f06d287330d8bf5ee eef9f32e75e5581c3561fcb17cd2dc21085761e6 M      test_apps
:040000 040000 ae9d9dfa93d8af9eb44623fd2b4466036a5f4cbb 154ad42f68f34472427c9bf1b6a07ff7f199882d M      tests

Comment 7

[Jon Hylands [:jhylands]]

This appears to be a duplucate of bug 987994... Alive, if it is, can you mark it as such?

Comment 8

[(inactive after 6/18) Alive Kuo [:alive]]

Dupe of 987994 and will be fixed in 950673 soon.