[Buildbot-devel] Buildbot is back (and reporting errors)

Niklaus Giger niklaus.giger at member.fsf.org
Sun Dec 23 15:48:30 UTC 2007


Hi

During the last two weeks I spent several evenings to bring the buildbot back 
into a good shape.

Please take a look at http://ngiger.dyndns.org/buildbot/ to see the some of 
the results.

I upgraded to the 0.7.6 buildbot revision which has a much improved web display.

It is therefore possible to limit one's view to specific branches (2.3.x/2.4.x/
2.5.x at the moment or just a specifiy builder (e.g. i386_f). I hope that this
makes looking for specific errors easier for everybody.

Most builder use a new approach. As before the buildbot checks out the desired
trunk/branch/tag, downloads the config and (if necessary) a board specific patch.
Then the attached make_xenomai.rb Ruby creates a script to checkout the desired
git branch, running the prepare-kernel.sh, configure, make etc. A short for is
available as in
http://ngiger.dyndns.org/buildbot/builders/yosemite_f/builds/0/steps/genscript/logs/cmd2run
The full script (which each line surrounded by a if ! cmd ; then exit 2; fi) is
then run. This approach should ease a developpers work to reproduce the environment
where the buildbot reported an error.

In the next few days (beside Christmas with my familiy, some skiing and friends)
I will improve it to handle "old" version like 2.3.x correctly. Also I will I
builds of the most recent tags in the 2.3. and 2.4 line. Or does someone see the
need for lines like 2.2/2.1 or 2.0?

Also planned is to add one ore more boards to running xeno-test, the Yosemite 
(PPC440EP based) and HCU3 (PPC405GPr based).

If somebody would like to add more boards please drop me an email (may be with
an attached config and a script to show how the build was done).

Now about the failures I noticed:

a) The simulator fails in two tests with the vxworks skins. In
http://ngiger.dyndns.org/buildbot/builders/sim_f/builds/1/steps/check_sim/logs/stdio
you will find at the end the lines
> Some problems were reported in:
> vxworks->t010726-1::
> vxworks->t010823-2::
Going back one finds:
> starting VxWorks services.
> Xenoscope: lt-t010823-2: fatal in MvmIrq (time=30000718.974563):
> test interrupted by watchdog.
and
> starting VxWorks services.
> Xenoscope: lt-t010726-1: fatal in MvmIrq (time=30000718.974820):
> test interrupted by watchdog.

b) The i386 builds fails in
> /home/buildbot/slave/i386_f/linux-2.6-xenomai/fs/jfs/jfs_metapage.c: In
> function 'metapage_read_end_io':
> /home/buildbot/slave/i386_f/linux-2.6-xenomai/fs/jfs/jfs_metapage.c:295:
> warning: 'return' with a value, in function returning void
> /home/buildbot/slave/i386_f/linux-2.6-xenomai/fs/jfs/jfs_metapage.c: In
> function 'metapage_write_end_io':
> /home/buildbot/slave/i386_f/linux-2.6-xenomai/fs/jfs/jfs_metapage.c:352:
> warning: 'return' with a value, in function returning void CC [M] 
> fs/jfs/jfs_logmgr.o
> /home/buildbot/slave/i386_f/linux-2.6-xenomai/fs/jfs/jfs_logmgr.c: In
> function 'lbmStartIO':
> /home/buildbot/slave/i386_f/linux-2.6-xenomai/fs/jfs/jfs_logmgr.c:2165:
> error: too many arguments to function 'lbmIODone'
> /home/buildbot/slave/i386_f/linux-2.6-xenomai/fs/jfs/jfs_logmgr.c: In
> function 'lbmIODone':
> /home/buildbot/slave/i386_f/linux-2.6-xenomai/fs/jfs/jfs_logmgr.c:2238:
> warning: 'return' with a value, in function returning void
> /home/buildbot/slave/i386_f/linux-2.6-xenomai/fs/jfs/jfs_logmgr.c:2263:
> warning: 'return' with a value, in function returning void
> /home/buildbot/slave/i386_f/linux-2.6-xenomai/fs/jfs/jfs_logmgr.c:2343:
> warning: 'return' with a value, in function returning void make[3]: ***
> [fs/jfs/jfs_logmgr.o] Error 1

-- 
Niklaus Giger
-------------- next part --------------
A non-text attachment was scrubbed...
Name: make_xenomai.rb
Type: application/x-ruby
Size: 11284 bytes
Desc: not available
URL: <http://buildbot.net/pipermail/devel/attachments/20071223/3699bfec/attachment.bin>


More information about the devel mailing list