Re: annotated bulk failures, pre-2011Q4 edition

[David Holland <dholland-pkgtech%netbsd.org@localhost>]

On Mon, Dec 19, 2011 at 03:49:26PM +0100, Thomas Klausner wrote:
 > On Mon, Dec 19, 2011 at 04:53:05AM +0000, David Holland wrote:
 > > audio/wsoundserver problems finding libaudiofile (even though buildlinked)
 > 
 > [...] Fixed.

thanks :-)

 > > benchmarks/nbench  (*) bad seddery
 > 
 > Fixed.

-    gccversion=`$compiler -v 2>&1 | sed -e "/version/!d"`
+    gccversion=`$compiler -v 2>&1 | sed -e "/ version/!d"`

So, any idea why it was working in my environment, which is also gcc
4.5? As far as I know it differs from Joerg's in only a handful of
ways:
   - two weeks' or so newer -current
   - base is built without Kerberos (this explains some of the krb
     divergence)
   - bootstrapped, which uses noticeably different <bsd.*.mk>
   - non-default $PREFIX and $VARBASE
   - not chrooted
   - no root availability

 > > cross/i386-netbsdpe        should just be removed
 > 
 > Do it :)

Done ;-)

 > > databases/cstore   (**) C++ issues
 > 
 > 1GB distfile, wow.

Yeah, this came up before; I think it's full of sample data or
something silly like that.

 > > devel/valide               requires newer vala
 > 
 > Someone please update vala, then...

It's worse than that; it requires a whole bunch of new stuff and
there's some thorny and intractable problem. obache was telling me a
while back that he tried and gave up in disgust

This is not something we should try to do right now I think. I guess I
should mark it BROKEN...

 > > devel/xsd          C++ issues
 > 
 > To be more exact, boost::filesystem::path problems. Also, xsd-3.3.0 is
 > out, perhaps that improves the situation (but it changed the build
 > system to GNU make).

Yeah, I thought I recognized that identifier but there weren't other
overt signs of boost so I wasn't sure...

-- 
David A. Holland
dholland%netbsd.org@localhost