Hi Danek,
On Wed, 2010-09-15 at 09:55 -0700, Danek Duvall wrote:
> Tim Foster wrote:
> > That makes sense alright, I've added that to this webrev: I have this so
> > that reference repository lint runs won't redirect stdout/stderr and
> > pkglint errors (not warnings) will cause the build to fail.
>
> I'm a bit confused. The new webrev now has pkglint run twice if
> PKGLINT_REF_REPO is defined. I would expect just the one run, against the
> reference repo
The intent was that the lint run against just the contents of our
repository, comparing whitelist output, would still be informative even
when PKGLINT_REF_REPO was defined. Given that it executes so quickly
compared to a full lint run, the extra step wasn't much overhead.
That said, I don't feel strongly about it, so it was no problem to have
it only do one lint run instead.
> I don't think you can have a blank line in a make rule (line 150).
dmake and usr/bin/make don't seem to mind them on a simple test
Makefile,
all:
echo "This is line one"
echo "This is line two"
but I've removed the blank lines nonetheless.
> > > ERROR opensolaris.action001.1 Username pkg$:::$::$:$:$:5srv in
> > > pkg://pkg5-nightly/package/pkg > 8 chars
> > > ERROR opensolaris.action001.3 Username pkg$:::$::$:$:$:5srv in
> > > pkg://pkg5-nightly/package/pkg is invalid - see passwd(4)
>
> It occurs to me that if we end up with expected errors that are no longer
> there, they should count only as warnings, and not build-breakers.
That makes sense. I've got a fix for this, and while the lint target is
getting a little lengthy, I think it's still readable (just)
Updated webrev at:
http://cr.opensolaris.org/~timf/pkglint-misc.hg
cheers,
tim
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
