Some patches randomly come back with those errors:

Count matched	Other	dblog.test	432	DBLogTestCase->testFilter()	
Count matched	Other	dblog.test	452	DBLogTestCase->testFilter()	
Count matched	Other	dblog.test	465	DBLogTestCase->testFilter()

Seen in http://qa.drupal.org/pifr/test/19520 (slave #34), and in some other issue I can't remember.

Let's try to reproduce.

CommentFileSizeAuthor
#1 dummy.patch524 bytesyched
dummy.patch524 bytesyched

Comments

yched’s picture

yched commented
StatusFileSize
new dummy.patch524 bytes

Try again.

deekayen’s picture

deekayen commented
Project: Drupal core » Project Issue File Review
Version: 7.x-dev » 6.x-2.x-dev
Component: dblog.module » Code

Those were both tested by #34 and passed. http://drupal.org/node/638356#comment-2287640 is a huge patch, but is it possible that MySQL could experience concurrency problems while testing like have been reported with the pgsql trials to explain the apparent inconsistency or does simpletest handle each test effectively in its own simpletest_watchdog table so they wouldn't interfere?

I told the clients to re-test themselves and they passed on a clean core test. #32 was in a failed state before I did so.

yched’s picture

yched commented
Title: [HEAD broken] Random failures in dblog tests » Random failures in dblog tests ?

Yes, it's awkward that I scream and yet cannot reproduce this on demand :-/. So let's quit screaming 'HEAD BROKEN' for now.
I'll try to have #638356: Reorganize field_test.module successfully retested.

I now remember what's the other patch where I saw this: http://drupal.org/node/606934#comment-2285338
Patch was *really* small, and had little chances of breaking anything. The failures were not reported again after I asked a re-test on the exact same patch.

I don't really understand what would cause this either, because DBLogTestCase::testFilter() does a db_delete('watchdog')->execute(); first thing, so there is little opportunity for more / less rows in the watchdog than what the test inserts itself.

deekayen’s picture

deekayen commented
Priority: Critical » Normal

If it's always a dblog failure, maybe the DBLogTestCase->testFilter() part needs some revision? If Jimmy thinks it's not a PIFR issue, then bounce it back to dblog.

yched’s picture

yched commented
Priority: Normal » Critical

FWIW, #638356: Reorganize field_test.module returned green on #33 (which insisted on finding PHP syntax errors earlier today - go figure...)

So, for the same patch file:
#33 5 hours ago: PHP syntax error - no log, crushed by the retest. http://qa.drupal.org/pifr/test/19506 is for a slightly different version of the patch.
#34 4 hours ago: 3 failures in dblog - http://qa.drupal.org/pifr/test/19520
#33 10 minutes ago: green - http://qa.drupal.org/pifr/test/19562

Poor patch got a bumpy afternoon ;-)

boombatower’s picture

boombatower commented
Project: Project Issue File Review » Drupal core
Version: 6.x-2.x-dev » 7.x-dev
Component: Code » dblog.module

This sounds like the other issues we have had with tests. Unless there is an issue with the bot itself...there isn't anything pifr is doing..other than invoking the tests...so this is all within simpletest and the test. Concurrency may play an issue, but once again that is part of simpletest.

sun.core’s picture

sun.core commented
Priority: Critical » Normal

Tests don't qualify as critical.

What's left here?

aspilicious’s picture

aspilicious commented

I never saw this failure (since my arrival in januar).
This is fixed???
Someone?

boombatower’s picture

boombatower commented
Status: Needs review » Fixed

Lets re-open if necessary.

Status: Fixed » Closed (fixed)

Automatically closed -- issue fixed for 2 weeks with no activity.