554970 - Socorro config parameter regression

Status: RESOLVED FIXED

(Socorro :: General, task)

Description

[K Lars Lohn [:lars] [:klohn]]

Recently, Frank made some changes to the way that the way that the materialized views handle configuration parameters in an attempt to simplify them.  Unfortunately, now it appears that they don't work at all for trying to regenerate data for a single product version.

The trouble is blocking testing for bug 534443 as I cannot find a combination of data parameters that will not result in an error.  What has worked in the past no longer works.

The configuration parameters are explained in the "Materialized Views Config" section of http://code.google.com/p/socorro/wiki/SocorroCommonConfig  These instructions are seemingly thorough, though they do not lead to getting something to work.

We one of two things: either the correct parameter combination to make it work or Frank's changes backed out to a point where what used to work, works again...

example that used to work: 

./startTopCrashesBySignature.py --product=Camino --version=2.0.3pre --startDate=2010-02-17 --endDate=2010-03-22T16:00 --startWindow=2010-02-17T00:00 --endWindow=2010-02-17T01:00 --deltaWindow=01:00

Comment 1

[K Lars Lohn [:lars] [:klohn]]

the error for the parameter combo that used to work:  

Traceback (most recent call last):
  File "./startTopCrashesBySignature.py", line 44, in ?
    tc = topcrasher.TopCrashesBySignature(configContext)
  File "/home/lars/breakpad/socorro/cron/topCrashesBySignature.py", line 69, in __init__
    self.startDate,self.deltaDate,self.endDate,self.startWindow, self.deltaWindow, self.endWindow = cron_util.getDateAndWindow(self.configContext,resultTable,self.productVersionRestriction,cursor,logger,**kwargs)
  File "/home/lars/breakpad/socorro/cron/util.py", line 440, in getDateAndWindow
    assert myConfig['deltaDate'] > datetime.timedelta(0)
TypeError: can't compare datetime.timedelta to NoneType

Comment 2

[Laura Thomson :laura]

Attachment: Back out r1859 while leaving in later unrelated changes to these files

I figure we should back out the broken code from trunk for now.  This patch reversed gives a starting point for actually fixing the issue

Comment 3

[K Lars Lohn [:lars] [:klohn]]

Comment on attachment 434952 [details] [diff] [review]
Back out r1859 while leaving in later unrelated changes to these files

I applied this to trunk and now I can successfully run my tests.

Comment 4

[Laura Thomson :laura]

In r1892.