gobject-introspection is like, the worst.

[maya%netbsd.org@localhost]

On Wed, Jun 28, 2017 at 02:20:36PM -0400, D'Arcy Cain wrote:
> On 06/28/17 13:11, maya%netbsd.org@localhost wrote:
> > On Wed, Jun 28, 2017 at 12:20:49PM -0400, D'Arcy Cain wrote:
> > > On 06/27/17 16:16, Joerg Sonnenberger wrote:
> > > > pkgsrc bulk build report
> > > > ========================
> > > > 
> > > > NetBSD 8.99.1/x86_64
> > > > Compiler: clang
> > > 
> > > I have not been able to build everything I need on -current.  Is it because
> > > I am still using gcc?  Has anyone else been able to get everything built
> > > using gcc or are we officially a clang OS now?
> > 
> > Joerg is a clang OS.
> > What are you having trouble building?
> 
> Lots of things.  They all fail in a similar way.  It fails running something
> called GISCAN and drops a Python core file that has this backtrace.
> 
> #0  0x000070e05a605663 in ?? ()
> #1  0x000070e05a200585 in ?? () from /usr/pkg/lib/libgthread-2.0.so.0
> #2  0x000070e0665e9b60 in ?? ()
> #3  0x000070e05a200669 in _fini () from /usr/pkg/lib/libgthread-2.0.so.0
> #4  0x0000000000000000 in ?? ()
> 
> I'm trying clang now.  It looks interesting in any case.

I don't run into this issue and I use GCC on NetBSD.

From the discussions on it, it seems a lot of the people who ran into
this issue are using pkg-rolling-replace and update builds. does a
completely fresh /usr/pkg /var/db/pkg also run into the same issue? (you
can temporarily mv them, build in a chroot, etc).

(Also, moving to a better list.)