David Baron :dbaron: Reporter
	Description • 16 years ago

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox3.5/1242921647.1242927642.4801.gz is a log of a unit test run that reported failure with no diagnostic that would suggest failure other than the "TinderboxPrint: mochitest-plain T-FAIL".  There should be diagnostic output in the log explaining why the test failed.  The lack of such output makes the failure difficult to diagnose and is holding the tree closed right now, although we'll probably sort that out somehow without this being fixed.

	David Baron :dbaron: Reporter
	Comment 1 • 16 years ago

To be clear, this failure has been happening for every Firefox3.5 unit test run since the *second* run with http://hg.mozilla.org/releases/mozilla-1.9.1/rev/525e43509e79 .

Might there have been an infrastructure change?

	David Baron :dbaron: Reporter
	Comment 2 • 16 years ago

Allegedly caused by bug 473259, which is being backed out.

	Mike Shaver (:shaver -- probably not reading bugmail closely)
	Comment 3 • 16 years ago

More specifically, they say they've _passed_ and then reporting a fail:

14:25 < mconnor> mochitest-chrome passed
14:25 < mconnor> program finished with exit code 0
14:25 < mconnor> elapsedTime=112.172000
14:25 < mconnor> TinderboxPrint: mochitest-chrome<br/><em 
                 class="testfail">T-FAIL</em>

	John O'Duinn [:joduinn] (please use "needinfo?" flag)
	Comment 4 • 16 years ago

Should this be DUP'd with bug#473259?

	Serge Gautherie (:sgautherie)
	Comment 5 • 16 years ago

(In reply to comment #0)
> unit test run that reported failure with no diagnostic that would
> suggest failure other than the "TinderboxPrint: mochitest-plain T-FAIL".  There
> should be diagnostic output in the log explaining why the test failed.

Actually, 'T-FAIL' means the count summary parsing failed.

(In reply to comment #1)
> Might there have been an infrastructure change?

In this case, this is due to a wrong count line format in the (checked in) patch :-<
My apologies.

(In reply to comment #2)
> Allegedly caused by bug 473259, which is being backed out.

The back out fixed this regression.