[Mono-devel-list] Corlib test cases

Nick Drochak ndrochak at gol.com
Fri Jan 23 06:14:02 EST 2004

|  -----Original Message-----
|  From: mono-devel-list-admin at lists.ximian.com [mailto:mono-devel-list-
|  admin at lists.ximian.com] On Behalf Of David Sheldon
|  Sent: Friday, January 23, 2004 8:44 AM
|  To: mono-devel-list at lists.ximian.com
|  Subject: Re: [Mono-devel-list] Corlib test cases
|  On Thu, Jan 22, 2004 at 10:05:10PM +0000, David Sheldon wrote:
|  > I have been trying to assign bug numbers to the tests that I can see
|  > failing.

Hi. I'm glad you are doing this!

The corlib tests all pass for me on windows (.NET 1.1).  So any failures
when run on mono (Linux or Windows) indicate at the least a compatibility
issue, and at worst an honest-to-goodness bug.

Note that many of the corlib tests have [Ignore] on them right now, 40 to be
exact, so there still might be some issues with those tests even on .NET.

I would say that we leave in all the tests, even if they don't make sense or
seem useless.  At least those give us an indication of where we might have
problems, and too many Asserts is not our biggest problem right now. My
thinking is if mono doesn't behave the same as .NET then we have a bug.  Of
course, true bugs in .NET are another issue.  Whether we have bug-for-bug
compatibility is a different topic :)

I started a kind of convention before where I'd enter a bug and then put an
[Ignore] on that test with a comment with the bug number in it.  Then once
bugs are closed we can re-enable those tests and make sure the bug really
was fixed.  If we could make this re-enable happen automagically, that would
be great.  Any ideas there?

Nick D.

More information about the Mono-devel-list mailing list