Hi Danek,

Thanks for taking a look.

On Thu, 2010-01-21 at 10:06 -0800, Danek Duvall wrote:
> Tim Foster wrote:
> 
> > http://cr.opensolaris.org/~timf/10748-webrev/
> 
> I think it would be useful to have REF_URIS not depend on REPO, since many
> people will want to resolve against the master repo, not the one they're
> publishing to.  Perhaps a REF_REPO which points by default to
> ipkg.sfbay/release?

Sounds good.

> Do we need the rule for "%/all_unbundleds" anymore?  Do we need the rule on
> line 675 anymore?

The first looks to be a convenience target, the second we don't need,
I'll remove it.

> Could all_unbundleds simply depend on both $(UNBUNDLED_TARGETS) and
> $(UNBUNDLED_CLUSTERS)?

Yes, I think so.  That said, while doing further testing, I've hit a
problem with the clusters import - the changes to the Makefile work fine
when importing svr4 packages, but don't work when trying to add the
cluster packages (java-dev, hpc-dev, etc.)

I was hitting a divide-by-zero traceback before fixing importer.py to
check for that, now with a simple test package 'timf-dev':

First pass: initial import 2010-01-22 03:06:07.903543
Package 'timf-dev'
  Version: 0.5.11-0.111
  Description: 
  Summary: Java Development cluster
  Classification:  org.opensolaris.category.2008:Development/Java
Second pass: global crosschecks 2010-01-22 03:06:07.911366
packages being published are self consistent
downloading and checking external references
INFO: ignoring action in [email protected]: unknown path=pkginfo 
variant.arch=i386
INFO: ignoring action in [email protected]: unknown path=pkginfo 
variant.arch=i386
INFO: ignoring action in [email protected]: unknown path=pkginfo 
variant.arch=i386
INFO: ignoring action in [email protected]: unknown path=pkginfo 
variant.arch=i386
INFO: ignoring action in [email protected]: unknown path=pkginfo 
variant.arch=i386
INFO: ignoring action in [email protected]: unknown path=pkginfo 
variant.arch=i386
external packages checked for conflicts
Third pass: dependency id, resolution and publication 2010-01-22 03:18:52.043297
0/0 packages processed
Done: 2010-01-22 03:18:52.043377
publication took 1:43

I'm still trying to track down what's going on here.

> Please be sure to use tabs to indent lines in the Makefile.

Sorry, yes.

> No need for the first hunk in importer.py, as all it does is add a bunch of
> spaces to the end of the line.

Oops, that change didn't show up in the sdiffs or cdiffs, only the wdiffs
and the patch (neither of which I looked at), I'll be more diligent
next time.

        cheers,
                        tim


_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss

Reply via email to