[OpenAFS-devel] Re: hudson success (i.e. automatically building)

[Jason Edgecombe]

On 07/13/2010 02:04 PM, Marcus Watts wrote:
>> Date:    Tue, 13 Jul 2010 11:01:51 CDT
>> To:      openafs-devel@openafs.org
>> From:    Andrew Deason<adeason@sinenomine.net>
>> Subject: [OpenAFS-devel] Re: hudson success (i.e. automatically building)
>>
>> On Mon, 12 Jul 2010 23:33:45 -0400
>> Marcus Watts<mdw@umich.edu>  wrote:
>>
>>      
>>> Other variations:
>>> heimdal vs. MIT. (vs. neither).
>>>
>>> --enable-demand-attach
>>> --enable-disconnected
>>> etc.
>>>        
>> Keep in mind many configure options are going to be going away soon.
>> That doesn't mean you shouldn't do this for now if you want... but you
>> don't want to be running a build with --enable-demand-attach _and_ one
>> without it in the next week or two, since you'll be doing twice the work
>> you need to.
>>
>> -- 
>> Andrew Deason
>> adeason@sinenomine.net
>>      
> I was assuming Jason Edgecombe had already seen that discussion.
> Obviously the list of supported configuration options will differ
> between versions.  I doubt that list will ever be empty.
>
> It would be nice to get the list of options from configure, then march
> through the list trying each one by one.  That may be way too complicated,
> especially with options that 'need something' to actually work.
> And this may not fit in well with the way hudson (or buildbot) see
> the world.
>
> Another thing to try, "make install" vs. "make dest" -- and whether the
> same set of things gets installed between the two.  In the same vein,
> "--enable-transarc-paths".  Perhaps some of this messiness will disappear
> when configuration options are simplified.
>    
Thanks to Marcus and everyone for their suggestions. Unfortunately, I'm 
still trying to get this thing crawling before walking or running. I 
think having a correct builder that fails or succeeds when appropriate 
is the first priority. Working with approved gerrit  submitters is the 
next priority.

Jason