Cyrus IMAPd 2.4.1 Released

[Greg Banks]

Bron Gondwana wrote:
> On Tue, Oct 19, 2010 at 08:41:51AM +0100, Jeroen van Meeuwen (Kolab Systems) wrote:
>   
>>     
>> Since we're on the subject...
>>
>> Please consider allowing me to deploy Hudson Continuous Integration.
>>     
>
> I heartily endorse this product and/or service.
>   

We used Hudson at my last job. Hudson rocks, assuming that the build 
(including any regression tests you run) is stable and predictable.  
It's very important to avoid spurious build failures, because that just 
gets the Hudson emails ignored, which takes away all its value.

>
>
> Greg is looking into testing as well.  I should put you two in contact!
>
>   

Hi Jeroen :)

I've been looking at the testing angle for three reasons:

a) it's sort of a personal fetish of mine anyway

b) some parts of the code are clearly untested, and I've only read a 
small fraction of it so far

c) my work for Fastmail will include doing significant surgery to some 
parts of the code, which worries me, given the absence of a significant 
regression test

In the short term I was hoping to get some point regression tests for 
some of the lower-level code happening.  I was thinking of using CUnit 
http://cunit.sourceforge.net/ to help manage the tests.  Thoughts?

After that I was hoping for some end-to-end regression tests using IMAP 
and POP3 clients written in Perl, and maybe do a test coverage study.  
Rob Mueller tells me there's no official Cyrus regression test suite 
currently, although he did point me at a Perl script he wrote back in 
2004 that might be a good starting point.  Thoughts?

-- 
Greg.