On 1 August 2012 23:28, Chris Cunnington <smalltalktelevision at gmail.com> wrote:
>
> /var/lib/jenkins
>
> drwxr-xr-x  2 jenkins     jenkins     4096 Aug  1 10:51 updates
> drwxr-xr-x  2 jenkins     jenkins     4096 Feb 29 09:51 userContent
> drwxr-xr-x  5 jenkins     jenkins     4096 Jul 20 12:49 users
> drwxr-xr-x  9 jenkins     jenkins     4096 Feb 29 09:50 war
> drwxr-xr-x  3 teamjenkins teamjenkins 4096 Jul 31 13:04 workspace
> -rw-r--r--  1 jenkins     jenkins        0 Aug  1 15:01 Workspace
> clean-up.log
>
> One of these things is not like the others. One of these things doesn't
> belong.
>
> drwxr-xr-x  2 jenkins jenkins 4096 Aug  1 10:51 updates
> drwxr-xr-x  2 jenkins jenkins 4096 Feb 29 09:51 userContent
> drwxr-xr-x  5 jenkins jenkins 4096 Jul 20 12:49 users
> drwxr-xr-x  9 jenkins jenkins 4096 Feb 29 09:50 war
> drwxr-xr-x  3 jenkins jenkins 4096 Jul 31 13:04 workspace
> -rw-r--r--  1 jenkins jenkins    0 Aug  1 15:01 Workspace clean-up.log
>
>
> Try it now.

OK, I think I've worked around the problem... by copying the job.
Hint: don't put spaces in the job name, because it looks like cog's
bin/squeak doesn't like spaces. I now have the test image firing up,
and it fails (rightly) because HDTestReport's not in the image. In
other words, tests.st must first load this before running the tests
themselves. Not a trainsmash. Progress, in fact. Can anyone beat me to
the punch and tell me from which Monticello repo I might grab such a
class? An Installer script lying around anywhere?

One thing that is screaming out at me is that we need to look at how
Pharo do their testing. IIRC, they had some really nice handling of
stdout. Right now I test the build script locally, headful and in the
foreground, so I can actually see what's happening. I'd really really
like the image to dump errors/stacktraces/etc. to stdout or stderr as
appropriate. That would be really really useful for the builds so we
can see what's actually going on.

I'm happy for someone to say "but here's how you can do it right now!".

frank