Re: package failures for 2012Q2

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

On Thu, Jun 14, 2012 at 12:04:54PM +0400, Aleksej Saushev wrote:
 > > devel/sdcc
 > >    This fails in some NetBSD environments, depending on the exact
 > >    pattern and length of pathnames, because it uses the find
 > >    -exec {} + syntax, which in NetBSD's find has completely broken
 > >    size accounting and sometimes fails with E2BIG.
 > 
 > Shouldn't this be fixed in NetBSD first?

Yes. I've been meaning to do that, but...

 > Also, I have found (on some linux system at work) that "-exec ... +" syntax
 > is not universally accepted, though I don't want to dive into details.

Hrm. Maybe the package should just be patched to not use it.

 > > lang/cint
 > >    Assorted C++ issues. Also annoying to work with because the
 > >    build doesn't stop on error.
 > 
 > It should be updated too.

ok, noted.

 > > lang/clang
 > >    Fails on netbsd-5 because -fPIC is not being used.
 > 
 > Perhaps needs update.

More likely someone needs to figure out why it's configuring
incorrectly.

 > > lang/racket
 > > lang/racket-textual
 > >    SIGABRT during build. Works on 32-bit platforms.
 > 
 > I'll try to bring it to upstream.

Thanks.

 > > math/octave-forge
 > >    Needs complete rework for current octave.
 > 
 > Octave should be updated too.

Yes indeed, but it's not broken yet :-)

 > > misc/root
 > >    Fails on fpos_t on -current and -6. Will probably have other
 > >    problems when that's fixed.
 > 
 > May need update.

I think there may be an update but I'm not at all convinced we can
expect that to fix it. It is related to cint in some fashion and has
roughly the same code quality.

 > > net/mppe-lkm
 > >    Requires kernel source; but even if the kernel source is
 > >    present, still doesn't build, at least on -current.
 > 
 > I suspect that it bitrotted way too much. Same as ndis(4).

Probably.

 > > parallel/openmpi
 > >    Various plugins don't install. Currently building an update.
 > 
 > This should be fixed already.

Yup.

 > > shells/ast-ksh
 > > shells/static-ast-ksh
 > >    Unfetchable. May be fixed.
 > 
 > Should be fixed already.

Yup.

 > static-ast-ksh may have further problems.

Indeed it does, and the behavior doesn't make much sense, but I get
the feeling it has a Special(TM) build system...

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