cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefano Mazzocchi <stef...@apache.org>
Subject [yes, you!] Gump is your friend!
Date Sat, 06 Nov 2004 17:31:16 GMT
giacomo@apache.org wrote:
> Author: giacomo
> Date: Sat Nov  6 04:38:55 2004
> New Revision: 56757
> 
> Modified:
>    cocoon/trunk/gump.xml
> Log:
> made concurrent-util dependency for core ??
> 
> Modified: cocoon/trunk/gump.xml
> ==============================================================================
> --- cocoon/trunk/gump.xml	(original)
> +++ cocoon/trunk/gump.xml	Sat Nov  6 04:38:55 2004
> @@ -83,6 +83,7 @@
>      <depend project="junit"/>
>      <depend project="jmock"/>
>      <depend project="eclipse" id="jdtcore"/>
> +    <depend project="concurrent"/>

thanks Giacomo, yeah, this will do it.

People, we've worked our collective ars off to make gump build cocoon 
(had to convince about a million people to do about a million different 
things) but gump moved from 65% (when I started pushing) to 92% as of 
yesterday (before giacomo reduced it to todays 85%).

In case you are not aware of how much things we are talking about, we 
are currently buildling:

   705!

different projects. Yes, you heard it: the current status of gump is 92% 
of 705 projects and given that a few cocoon blocks prevent other 10 
blocks to build, I expect to reach 95% when we are finished integrating.

For a more detailed description look at:

  http://brutus.apache.org/gump/public/buildLog.html

or, in case it's building

  http://brutus.apache.org/gump/test/buildLog.html

which runs only once a day and it's more likely to be finished.

Now.

Here are a few suggestions for people:

  1) if you don't know how to map a java package to the gump name of the 
project, look at this page

  http://brutus.apache.org/gump/public/gump_xref/package_project.html

it lists the java packages contained in each project and gives you a 
first rough indication of where to look for (and to avoid misnaming a 
project).

  2) if you find a problem down the road, or if you want to add a new 
project to gump, the Gump CVS module contains the metadata used by 
brutus to make the gump runs. That module is *open* to every ASF 
committer, so if you spot a problem, go right ahead and fix it. Don't be 
afraid of breaking things, we can always roll back and it takes a few 
hours for somebody to notice if things went wrong. We break gump all the 
time so it's not a big deal and nobody will shout at you ;-)

  3) if you care about non-standard builds, take a look at

  http://brutus.apache.org/gump/kaffe/

and

  http://brutus.apache.org/gump/jdk1.5/

these are new gump runs that we are attempting to build.

  4) I *KNOW* those pages are horrible and those email are noisy and 
full of crap that makes it harder to spot the problem. I'm working on it.

  5) I *KNOW* it's annoying that gump does not backtracks with the 
previous builds to find out who is really the project that generated the 
changes. I'll be working on that right after so that not only gump will 
be able to indicate who caused what effect but also gump will work 
*both* as a continuous integration system and as a nightly build system.

If you have any questions, suggestions, comments or criticism, I (and 
the other gumpmeisters) are very happy to receive them.

-- 
Stefano.


Mime
View raw message