Bug 124365

Summary:	at regression test fails 7 days before the end of each month
Product:	Red Hat Enterprise Linux 3	Reporter:	Charlie Bennett <ccb>
Component:	at	Assignee:	Jason Vas Dias <jvdias>
Status:	CLOSED DUPLICATE	QA Contact:	Mike McLean <mikem>
Severity:	medium	Docs Contact:
Priority:	medium
Version:	3.0	CC:	shillman
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2006-02-21 19:03:46 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Charlie Bennett 2004-05-25 23:23:30 UTC

Description of problem:

Every month(!) if today + 1 week carries us into next month.

It's the 'next week + 1 month' and 'next week - 1 month' tests.
parsetest gets it right, the perl code in the regression test
gets it wrong.  I'll drop you a line when I have a patch.

It's failing today.  Succeeded yesterday and tomorrow.  Will
fail again on the 24th of June....


Version-Release number of selected component (if applicable):

This may affect this package in other releases.

How reproducible:

Like clockwork.

Steps to Reproduce:
1. set your system clock to May 25 or June 24 or July 25, etc.
2. run make test in the at build directory
3.
  
Actual results:

tests of 'next week + 1 month' and 'next week - 1 month'
fail routinely.  tests of 'next week - 10 months' and
'next week + 10 months' fail for some months.

Expected results:

We expect the test harness to do at least as good at this
computation as the program it is testing.

Additional info:

The program under test gets it right.  The test harness is
the broken element.

I won't be moving forward tonight without generating a patch.
I'll post it here as soon as it's available.

Comment 2 Jens Petersen 2004-06-28 15:58:07 UTC

This is a duplicate of bug 108701 for RHEL3, but is still open
for FC2 and later I'm afraid.

Comment 3 Jason Vas Dias 2004-08-03 23:03:28 UTC

And for the latest at-3.1.8-56+ (with -t option).

I had to temporarily disable the 'make test' with  
  %define _without_check 1
to get it to build on July 29th - I wish I'd seen
this bug then!

It is on my to-do list to fix this, and I'll make 
a patch to 'test.pl' - this is not trivial.

At least I can confirm that at-3.1.8-57 passes the
test on August 3rd!

Comment 4 Jason Vas Dias 2004-08-04 23:01:05 UTC


*** This bug has been marked as a duplicate of 108701 ***

Comment 5 Red Hat Bugzilla 2006-02-21 19:03:46 UTC

Changed to 'CLOSED' state since 'RESOLVED' has been deprecated.