Alexey Verkhovsky wrote:
> On Jan 27, 2008 4:01 PM, M. Edward (Ed) Borasky <znmeb / cesmail.net> wrote:
>> Charles Oliver Nutter wrote:
>>> Perhaps it's time to set up a continuous integration server for ruby
>>> core that performs a number of tasks on commit or periodically:
>>> Time for a change?
>> Well ... it's a great idea.
> 
>>From my (mixed) experience with running a CI build for Rails [1] for
> the last few months, it only makes sense if the project team cares
> deeply enough about keeping the CI build green. Otherwise, it's
> becomes just an expensive source of permanent annoyance in everyone's
> mailbox.

So... the trick is to only include packages for which there are 
identified people with the interest and bandwidth to keep on top of the 
issues found.

I don't have the knowledge or expertise necessary to keep ruby core 
building successfully, but if we can get to the expectation that it 
generally builds and tests successfully, I could occasionally contribute 
and could definitely help isolate issues down to a specific change.  My 
last successful build was 15241; if we could get back on track I can 
generally identify the cause of a regression down to a specific checkin, 
and often down to a specific file.

For all of the other packages on my list, I do have the bandwidth, 
interest, and access to either directly make the fix, or the willingness 
to maintain a separate set of patches.  Once we collectively can get on 
top of those packages, I would be willing to expand the list of packages 
that I am looking out for.

Hopefully we can find others with an interest in contributing.

I see Ruby 1.9 in a similar state as a number of Java packages at the 
ASF were at in 1999, when I created Gump:

http://gump.apache.org/why.html

- Sam Ruby