Looking at the generated build.sh file, the classpath is set up correctly, and no variables are passed on the ant command that would explain this. I've grepped the various .properties and .xml files in the cocoon directory and don't see how the excalibur jars are referenced.
hmmm
Stefano, if you could take a look at this, I'd appreciate it.
oh no! Sam ruby is playing the jedy trick with me... gosh, I knew this would have happened one day or another. And he knows that if I dive into gump I'll be stuck there forever. Gosh.
Chuckle. I certainly wouldn't be above that. But in this case, I'm suggesting exactly the opposite. By snipping a few lines out of the generated build.sh you can actually remove Gump from the picture. And what I am asserting is that if Gump is removed, the failure will remain.
You don't actually have to go through the process of removing everything else from the generated build.sh to demonstrate this (you can if you like, but you don't *have* to). What I am asserting is that the there is no gump code running, nor is there any environment variables set by gump that would have caused this problem. This leads me to believe that the problem is in the build.xml, or in the various xml files that it references, or in the property files.
No need to look into the xslt files... simply load build.sh into your favorite editor and look for "cocoon;" (with the semicolon). Immediately after this you will see the classpath being set up and then the ant call. Tell me is you see anything amis.
All right, all right, I know I'm going to regret this, but I will :)
Again, my goal is actually not to overload you at this point. Just to provide you with enough information to solve this particular problem. The "sucking in" will happen later, I promise. ;-)
Even if this were corrected, however, I do see a jar in the avalon-excalibur/compatibility/build/lib directory. What's odd is that it was built this morning, but has yesterday's date on it. This I can't explain either.
are you sure the cron jobs are acting regularely?
Last night, while you slept but it was still reasonably early for me here, the cron job ran regularly. It failed spectacularly as there was a failure in Xalan and pretty much everything downstream was held up due to prereq. I corrected the Xalan problem (xsltc is now part of the build, so its prereqs needed to be added to the project definition). I then manually started a full build and went to bed.
All of this occurred on 2003-04-02, local time on the machine. And was done *after* the full build. So why there is any jars present with yesterday's date is beyond me.
Stefano.
- Sam Ruby
