885784 - disable mozdevice tests

Status: RESOLVED FIXED

(Testing :: Mozbase, defect)

Description

[Jeff Hammel]

https://tbpl.mozilla.org/php/getParsedLog.php?id=24427195&tree=Mozilla-Central

Well, our tests are unreliable.  Big surprise.  Lets disable the
beasts unless we want to resolve bug 790765 as incomplete or what not.

Comment 1

[Jeff Hammel]

Attachment: disable the bastards

to be landed on mozbase and m-c.  the patch is against mozbase;  if a separate patch/review is needed for m-c, lemme know

Comment 2

[Ed Morley [:emorley]]

https://tbpl.mozilla.org/php/getParsedLog.php?id=24427195&tree=Mozilla-Central

Comment 3

[Ed Morley [:emorley]]

err that was in comment 0 anyway, oh well :-)

Comment 4

[Jeff Hammel]

irc://irc.mozilla.org/#ateam

10:15 < jhammel> jgriffin, ahal, mcote : so inbound is closed and i'm on a 
                 train and will be off soon
10:15 < jhammel> i.e. i can't/shouldn't push currently
10:16 <@jgriffin> jhammel: ok, I'll push it later
10:16 < jhammel> jgriffin: thanks

Comment 5

[Jeff Hammel]

Obviously this is a complete sledgehammer, but I resorted to this for a few reasons:

- not sure if this is the only test that fails; assuming it is, we can just disable that test like civilized people

- instead of commenting out the inclusion, it should look more like:

"""
[incude:mozdevice/tests/manifest.ini]
disabled = https://bugzilla.mozilla.org/show_bug.cgi?id=885784
"""

However, there's no good expedient way of ensuring that all subsets are disabled...I think. Lemme comment on that later.

- since make check is a critical path of failure, this needs to be disabled now.  in emergencies, i tend to resort to sledgehammers that are assured success vs a researched solution

Comment 6

[Jonathan Griffin (:jgriffin)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/def4320f90aa

Comment 7

[Jonathan Griffin (:jgriffin)]

and github: https://github.com/mozilla/mozbase/commit/b7a939962fe06499258a45abaaa11007955593c1

Comment 8

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/mozilla-central/rev/def4320f90aa

Comment 9

[Jeff Hammel]

(In reply to Jeff Hammel [:jhammel] from comment #5)
<snip/>
> - instead of commenting out the inclusion, it should look more like:
> 
> """
> [incude:mozdevice/tests/manifest.ini]
> disabled = https://bugzilla.mozilla.org/show_bug.cgi?id=885784
> """
> 
> However, there's no good expedient way of ensuring that all subsets are
> disabled...I think. Lemme comment on that later.

Reading the code, this is in fact untrue:
https://github.com/mozilla/mozbase/blob/master/manifestdestiny/manifestparser/manifestparser.py#L693

Perhasp worth documenting.  In short, 'disabled = reason' wins.