774396 - B2G reftest-content.js fails to instantiate nsIPrintSettings object

Status: RESOLVED WONTFIX

(Testing :: Reftest, defect)

Description

[Andrew Halberstadt [:ahal]]

Attachment: reftest-sanity log

Between 10-14 tests (depending on run) are failing when running layout/reftests/reftest-sanity/reftest.list on the b2g emulator (see bug 770490).

If the cause of these failures is discovered and/or fixed then it may go a long way to fixing (or at least understanding) many of the failing reftests in the other manifests as well.

Reftest log attached.

Comment 1

[Andrew Halberstadt [:ahal]]

Attachment: reftest-sanity log

The previous log was from my laptop which seemed to have problems running the reftests. Now using a dedicated linux desktop machine (and with a couple bug fixes) things have settled down. I'm only seeing three failures now that are consistent. Updating the attached log.

Comment 2

[Andrew Halberstadt [:ahal]]

These are now running on cedar and I see the same three failures: https://tbpl.mozilla.org/php/getParsedLog.php?id=16666888&tree=Cedar

(Note that the other two failures in chunk 1 are due to bug 785074)

Comment 3

[Andrew Halberstadt [:ahal]]

:ted pointed out that all three of these bugs are testing printing. I remember when I first got this running, reftest-content.js couldn't instantiate an nsIPrintSettings object. This was back in June/July so maybe it will be fixed by now. I'll take a look into it when I have a bit of time.

Comment 4

[Jonathan Griffin (:jgriffin)]

It's quite possible that the nsIPrintSettings interface doesn't exist for B2G, in which case we should just skip these for B2G.

Comment 5

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

We may be able to use a dummy implementation.  I'm not sure what android does for this, but that's a good first place to look.

Comment 6

[Andrew Overholt [:overholt]]

I spoke with Andrew on IRC about this and he said we don't need to block on it.

Comment 7

[Andrew Halberstadt [:ahal]]

(In reply to Andrew Overholt [:overholt] from comment #6)
> I spoke with Andrew on IRC about this and he said we don't need to block on
> it.

AIUI this only affects printing tests which are of low importance to b2g.