Steve,
I was looking at the three builds on which sgmlkit is skipped (that's
TRU64-ALPHA, DEBIAN-3.0r3_i386 and SPARC-SOL9), and noticed that on
each, the 'rpms' target was marked 'skipped', presumably because the
sgmlkit is listed as 'failed'. Since rpms seems to be failing
persistently on RHEL-WS3-3_i386, there seems at least a chance that it
will have problems on the other three as well, but they won't be tested
as long as the nightly build script's problem with suppressed builds
still exists.
So, once again, that's:
test -f STAR_SUPPRESS_BUILD || make dist
instead of plain 'make dist'. This has been wrong for three months,
now: would it be terribly inconvenient to make the one fix in one line
of one file and stop seeing all these spurious sgmlkit failures?
Apart from anything else, it would be nice to be able to go to Kyoto
and say to the AstroGrid folk that we do in fact have working nightly
builds. We're _so_ close. As it turns out, we haven't ever actually
had working nightly builds for the classic applications (as opposed to
the Java ones), whereas the AG folk reportedly have had them for their
stuff since September. Not that it matters, of course, but it would be
nice.
All the best,
Norman
--
----------------------------------------------------------------------
Norman Gray : Physics & Astronomy, Glasgow University, UK
http://www.astro.gla.ac.uk/users/norman/ : www.starlink.ac.uk
|