[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [Libguestfs] [libnbd PATCH 0/5] Introduce automated testing using GitLab CI



On Tue, May 11, 2021 at 02:13:47PM +0100, Richard W.M. Jones wrote:
On Tue, May 11, 2021 at 01:56:14PM +0100, Daniel P. Berrangé wrote:
On Tue, May 11, 2021 at 01:34:19PM +0100, Richard W.M. Jones wrote:
> On Tue, May 11, 2021 at 12:16:19PM +0100, Richard W.M. Jones wrote:
> > On Thu, May 06, 2021 at 01:30:05PM +0200, Martin Kletzander wrote:
> > > FreeBSD
> > > =======
> > >
> > > The build fails because there is no fallocate() there.
> >
> > I'm actually not seeing a failure (FreeBSD 12.2).  Do you happen to
> > know what the build error was?
> >
> > I'm seeing test failures though which I'll fix if they are easy.  I
> > think for non-Linux right now I'm most concerned about build failures.
>
> Another problem with the libnbd tests is that they use the installed
> nbdkit, but -- presumably for political rather than technical reasons --
> nbdkit on FreeBSD is built without GnuTLS support.  While I could add
> 'requires' lines everywhere to test this, it may be better (on
> FreeBSD) to either not test against nbdkit at all or to test against
> the latest nbdkit build.  Could CI tests allow this?  Don't worry
> about implementing this, I just want to know what's possible.
>
> The situation for testing nbdkit is similar but reversed.  FreeBSD
> does not package libnbd at all so far, but I'm quite sure they'll also
> build a TLS-less version if they get around to it.

I wouldn't assume it is intentionale, as opposed to just a mistake.
They happily build other stuff with GNUTLS including both libvirt
and QEMU

Assuming this is current:

  https://cgit.freebsd.org/ports/plain/net/nbdkit/Makefile

it seems to suggest they're attempting to build with gnutls.

Perhaps the configure script failed to detect gnutls despite being
present and silently disabled it in their builds ?

Oh this is ports versus packages again.  Who thought haven't
two different packaging systems was a good idea?!

 $ nbdkit --dump-config
 bindir=/usr/local/bin
 filterdir=/usr/local/lib/nbdkit/filters
 libdir=/usr/local/lib
 mandir=/usr/local/man
 name=nbdkit
 plugindir=/usr/local/lib/nbdkit/plugins
 root_tls_certificates_dir=/usr/local/etc/pki/nbdkit
 sbindir=/usr/local/sbin
 selinux=no
 sysconfdir=/usr/local/etc
 tls=no                           <------
 version=1.20.4
 version_major=1
 version_minor=20

But this is the pkg version, not the ports version.


OK, I must admit I am not used to *BSD packaging and I am glad I got the
CI running on FreeBSD at least.

Somehow the ports version of nbdkit is not available on my FreeBSD
box.  I'm not sure if I'm doing something wrong or if it's because I'm
not using the latest nbdkit.  However I did read the link above and I
agree it ought to be using gnutls.

The bigger problem is installing a port in a CI test is I suppose more
likely to go wrong (because it's actually building the prerequisite
from source).


Well, if it fails, that is.  I guess we _could_ install from ports, at
least for some particular package.

@Andrea: Did you stumble upon anything related where you needed to deal
         with or thought about using something from ports?

Martin

Rich.

Regards,
Daniel
--
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|

--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-p2v converts physical machines to virtual machines.  Boot with a
live CD or over the network (PXE) and turn machines into KVM guests.
http://libguestfs.org/virt-v2v

Attachment: signature.asc
Description: PGP signature


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]