Re: py-scipy problem (from Re: Automated pkgsrc commit testing)

[Jason Bacon <jtocino%gmx.com@localhost>]

On 9/14/24 07:42, Jason Bacon wrote:
> On 9/13/24 16:56, Thomas Klausner wrote:
> > On Fri, Sep 13, 2024 at 04:49:45PM -0500, Jason Bacon wrote:
> > > I'm assuming this issue has been fixed by now.
> >
> > Yes.
> >
> > > I'm seeing failures on NetBSD 10, which I would guess are a new problem.
> > >   Anyone else having trouble with scipy on NetBSD?
> >
> > It builds in the official NetBSD 10 bulk builds, last two days ago.
> >
> > https://releng.netbsd.org/bulktracker/math/py-scipy
> >
> > list all of these:
> >
> > py311-scipy-1.14.1    ok    2024-09-12    HEAD    NetBSD 9.0/ 
> > i386        gcc
> > py311-scipy-1.14.1    ok    2024-09-12    HEAD    NetBSD 9.0/x86_64    
> > gcc
> > py311-scipy-1.14.1    ok    2024-09-11    HEAD    NetBSD 10.0/ 
> > x86_64    gcc
> > py311-scipy-1.14.1    ok    2024-09-11    HEAD    NetBSD 10.0/i386    gcc
> >
> > This is py311, but I checked the 10.0/x64_64 one, and the py312 
> > version builds too, see
> >
> > https://releng.netbsd.org/bulktracker/build/1246/math/
> >
> > Make sure all your dependencies are up to date?
> >   Thomas
>
> OK, probably a dependency update issue.  I'll raze my tree and try again...
>
> --
> Life is a game.  Play hard.  Play fair.  Have fun.

It looks like the presence of some other package is causing the build to 
fail.  I removed all packages (except the non-removable core packages), 
began reinstalling everything, and py-scipy failed again.  I then 
removed all packages again, and installed py-scipy before anything else, 
and it succeeded.

Since package builds are succeeding, I'm not going to worry about this 
for now, but if/when I encounter another failure, I'll try to identify 
the offending package.


--
Life is a game.  Play hard.  Play fair.  Have fun.