forrest-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Steven Noels <stev...@outerthought.org>
Subject Re: Forresbot WANTED
Date Wed, 27 Nov 2002 12:42:57 GMT
Sam Ruby wrote:

 > Steven Noels wrote:
 >>
 >> Wouldn't it be possible to just build the documentation instead?
 >> I'm very reluctant to have source code compilation running on that
 >> machine
 >
 > The proposal would be that gump would start running at approximately
 >  0200 GMT every day, and largely be dormant otherwise.

Can you provide me with some more data: required diskspace, on what 
hardware Gump is currently running, average run length, etc etc

 > The reason I am volunteering for this is that there essentially is no
 >  invention required on my part... All I will be doing is cloning
 > something I already have running on three machines.

Of course. This has nothing to do with trust, just with checking whether 
the machine is up to the task. And I'd prefer to check this upfront, 
rather than complain afterwards.

Diskspace is no issue, however:

Filesystem            Size  Used Avail Use% Mounted on
/dev/md2               77G   12G   61G  16% /
/dev/md1               20M   11M  9.1M  53% /boot

It's the CPU cycles I'm worried about - and the fact that a number of 
Tomcat JVMs are running in parallel, preferably with higher priority.

 > Nicola Ken Barozzi wrote:
 >
 >> Forrestbot has a major shortcoming: it uses sources from CVS and
 >> does not include stuff like javadocs and any doc that the
 >> application generates during compilation.

... which aren't integrated in any way with Forrest anyhow, and the only 
way we have been thinking about possible integration is through 
Ant-copying and not through Cocoon components. I need a more convincing 
case.

 > +1.
 >
 > The more information Forrest has access to, the better job it can do.

If this requires Forrest to pull all this information to a central 
location, we know that there will be days which will be simply too short 
to process all available information within a 24 hour timelapse.

I heared about execution times of multiple hours, and my relunctancy is 
partly based on the fact that I'd rather donate CPU cycles to live 
Cocoon hosting of cocoon.apache.org, rather than compiling all 
sourcecode of name-your-favourite-project only to generate its website.

I don't know much about Gump - would it be possible to create a special 
'profile' that only does Forrest generation, possibly with a list of 
other required doco-related resources like Javadoc?

We could of course turn the world around and have our Forrestbot depend 
on the gump profiles of the different projects.

</Steven>
-- 
Steven Noels                            http://outerthought.org/
Outerthought - Open Source, Java & XML Competence Support Center
Read my weblog at              http://radio.weblogs.com/0103539/
stevenn at outerthought.org                stevenn at apache.org


Mime
View raw message