launchpad-dev team mailing list archive

[Michael Hudson <michael.hudson@xxxxxxxxxxxxx>]

Julian Edwards wrote:
> Michael Nelson wrote:
>>
>> On Mon, Nov 23, 2009 at 8:57 AM, Michael Nelson
>> <michael.nelson@xxxxxxxxxxxxx <mailto:michael.nelson@xxxxxxxxxxxxx>> wrote:
>>
>>
>>
>>     On Wed, Nov 18, 2009 at 7:15 AM, Michael Hudson
>>     <michael.hudson@xxxxxxxxxxxxx <mailto:michael.hudson@xxxxxxxxxxxxx>>
>>     wrote:
>>
>>
>>         Julian Edwards wrote:
>>         > There is one nightmare part to this - the implementation of
>>         > IBuilder.findBuildCandidate().  This is currently a very ugly
>>         query that
>>         > decides which job to dispatch next, based on a few criteria
>>         about the jobs.  I
>>         > expect it to get to troll-like ugliness by the time we finish
>>         and I'm not sure
>>         > how it will work yet, it depends on how Michael re-factors
>>         IBuilder and sets
>>         > up an interface that other builders must implement.
>>
>>         Wow, that's a beauty alright.
>>
>>
>>     Hmm... as far as I understand it, we aren't going to have
>>     generalised builders that can do both binary-package building and
>>     branch->source package building initially right? Rather, we'll have
>>     different types of builders. So although we'll want to keep the
>>     scoring consistent for both so that when we *do* have such
>>     infrastructure they'll be queued fairly, currently each builder type
>>     will be able to implement its own findBuildCandidate() and only be
>>     looking for builds of the required type... in which case, the
>>     short-term solution for findBuildCandidate() should be pretty
>>     straight forward?
>>
>>
>> Erm - my mistake. We won't yet have generalised builders that can accept
>> jobs for different archs, but any builder should be able to do branch
>> jobs (ie. non soyuz-build jobs).
> 
> 
> I hope that's the case.  I guess it means the other job types need to be
>  script-based, which I think they are at least for the recipe stuff.

Well, the recipe stuff doesn't exist yet :-)  But I think it's a very
sensible idea to run possibly-unruly code in a subprocess and so I'm
sure we'll have a script as an entry point.

Cheers,
mwh