Exit status ignored by the snakefarm?

Post by Per Cederqvist

Build test succeded. Any warnings are appended below.
--
_______________________________________________
Snake-farm-report mailing list
http://lists.lysator.liu.se/mailman/listinfo/snake-farm-report

Hmm. I kill a couple of python processes, and the snake farm reports
success? Something doesn't take care of the exit status the way it
should.

Catching the exit status is difficult from make in this line
(buildit.sh, 146):

(make $target 9>&1 1>&2 2>&9 | tee $makeerrlog) > $makelog 2>&1

Post by Per Cederqvist
A few reports were sent to the snake-farm list. Those jobs were
started before the snake-farm-report list was created. There are
still 5 old jobs running, whose report might end up here sooner or
later.

Previously, they died when trying to gobble memory. Something has
changed. I'll get down to debugging them immediately:

As it turns out, test_import baloons the python interpreter from ~12MB
of RAM to ~42MB on fafner. corresponding manuever on linux-intel gives
only a modest increase. Needless to say, the general bloating makes
the interpreter hit the ulimit a bit later (config.fafner):

ulimit -d 65000
ulimit -s 8192
ulimit -t 1000
ulimit -v 65000

Unfortunately, regrtest -v on test_import doesn't give any extra
debugging output.

--
Anders "Quest" Qvist

"We've all heard that a million monkeys banging on a million typewriters
will eventually reproduce the entire works of Shakespeare. Now, thanks
to the Internet, we know this is not true." -- Robert Wilensky

M.-A. Lemburg

2002-07-24 15:33:33 UTC

Post by Per Cederqvist

Build test succeded. Any warnings are appended below.
--

Hmm. I kill a couple of python processes, and the snake farm reports
success? Something doesn't take care of the exit status the way it
should.

Catching the exit status is difficult from make in this line
(make $target 9>&1 1>&2 2>&9 | tee $makeerrlog) > $makelog 2>&1

Just curious: why don't you write the build code in Python ?

It's really easy to setup things like this by using the popen2
module.

--
Marc-Andre Lemburg
CEO eGenix.com Software GmbH
_______________________________________________________________________
eGenix.com -- Makers of the Python mx Extensions: mxDateTime,mxODBC,...
Python Consulting: http://www.egenix.com/
Python Software: http://www.egenix.com/files/python/

Anders Qvist

2002-07-24 15:58:13 UTC

Post by Per Cederqvist

Build test succeded. Any warnings are appended below.
--

Hmm. I kill a couple of python processes, and the snake farm reports
success? Something doesn't take care of the exit status the way it
should.

Catching the exit status is difficult from make in this line
(make $target 9>&1 1>&2 2>&9 | tee $makeerrlog) > $makelog 2>&1

Just curious: why don't you write the build code in Python ?
It's really easy to setup things like this by using the popen2
module.

If the build scripts were in Python, they wouldn't run where Python
don't. Eg Python don't easily compile on UNICOS. By being able to try
to build Python, we can wait for patches in CVS to result in a working
build.

M.-A. Lemburg

2002-07-24 16:51:51 UTC

Post by Anders Qvist
Catching the exit status is difficult from make in this line

(make $target 9>&1 1>&2 2>&9 | tee $makeerrlog) > $makelog 2>&1

Just curious: why don't you write the build code in Python ?
It's really easy to setup things like this by using the popen2
module.

Point taken. Just thought it would make the build script development
a little easier for you.

Perhaps you could use an older version such as Python 1.4 or
1.5.2 for the build process ? You wouldn't need things like
sockets which usually cause most of the trouble, so there's
a chance...

Anders Qvist

2002-07-24 20:04:18 UTC