[Gt-eos] GCT build system
Mischa Salle
msalle at nikhef.nl
Fri Jan 5 11:05:55 CET 2018
Hi Brian, Mátyás,
good news, thanks for all the hard work!!
A few small remarks:
- I noticed that we will need to change the spec files, they now still
contain e.g.
[...]
Vendor : Globus Support
URL : http://toolkit.globus.org/
Summary : Globus Toolkit - Globus GSI Proxy Core Library
Description :
The Globus Toolkit is an open source software toolkit used for
building Grid systems and applications. It is being developed by the
Globus Alliance and many others all over the world. A growing number
of projects and companies are using the Globus Toolkit to unlock the
potential of grids for their cause.
- concerning binary RPMs, copr is very straightforward, and easily
scriptable (I have just started a build using cURL for a test project,
see attached 2 scripts, plus the content of
https://copr.fedorainfracloud.org/api/, I also use the OBS for addons
for OpenSUSE, but actually find copr even easier). It would be
straightforward to run this for all the new src RPMs by making a
new gct project or so. It looks like it's possible to trigger the copr
build directly from git tags.
For a prerelease it doesn't make so much difference, especially if we
never will build our own production binary RPMs, there isn't much
point in spending effort getting it right, since for EPEL/Fedora we
will only need the src RPMs.
Mischa
On Thu, Jan 04, 2018 at 11:18:39AM -0600, Brian Lin wrote:
> Hi all,
>
> Stealing Mat's thunder since he's out of the office -- we've got pre-release
> tarballs and SRPMs available
> (https://github.com/gridcf/gct/releases/tag/v0.20180103-pre)! At this point,
> I think the next step is to get binary RPMs based off of the pre-release so
> we can test them, perhaps in EPEL?
>
> Regarding distribution of binary RPMs, copr/opensuse build services seems
> like overkill if we can get them into EPEL, imo.
>
> - Brian
>
> On 12/22/2017 12:24 PM, Dave Dykstra wrote:
> > I use the OpenSUSE Build Service for making signed centos yum
> > repositories and debian/ubuntu apt repositories
> > https://build.opensuse.org/repositories/home:cvmfs:contrib
> >
> > It reads all the source from github, and I have heard that builds can be
> > automated based on github triggers, but I haven't tried that yet.
> > Instead so far I manually update the configuration when it is time for a
> > new release.
> >
> > Dave
> >
> > On Thu, Dec 21, 2017 at 03:36:30PM +0100, Mischa Salle wrote:
> > > Hi Mátyás,
> > >
> > > it's for Mattias Ellert to further decide what to do, but in any case,
> > > it makes sense to me to only make src, not binary RPMs.
> > >
> > > Personally I'm not sure I would like to have binary RPMs directly made
> > > by travis. I don't know whether it could produce signed RPMs from it,
> > > a koji-based approach seems more reasonable. Perhaps we can use
> > > https://copr.fedorainfracloud.org/
> > > For 'our' RPMs we could also use the open(suse) build service, which
> > > produces signed RPMs, but copr seems more logical.
> > > For debs, I would say it might be better to just use the src
> > > distribution tarball, since you typically also want to sign the src
> > > packages, and do a manual update of the changelog.
> > >
> > > But perhaps Mattias has different ideas?
> > >
> > > Cheers,
> > > Mischa
> > >
> > > On Wed, Dec 20, 2017 at 03:45:30PM -0600, Mátyás Selmeci wrote:
> > > > Hi folks,
> > > >
> > > > Just to let you know where we are in regards to getting a build system
> > > > going. I???ve put together a PR for automated builds (using Travis-CI) such
> > > > that whenever a GitHub Release is tagged, automated tests are kicked off
> > > > and, if successful, tarballs and source RPMs for the individual components
> > > > are added to the download page for the release.
> > > >
> > > > I have not included binary RPMs because
> > > >
> > > > a) we???re probably going to get them from EPEL anyway;
> > > > b) deployment is unreliable when you have too many files in a release;
> > > > c) without a YUM repo, downloading and installing those RPMs would be too
> > > > painful anyway.
> > > >
> > > > I have provided a script and a Vagrantfile in case anyone does want to build
> > > > the binary RPMs for themselves. I haven???t done anything with debs because I
> > > > have no experience with Debian packaging, but if someone wants to put
> > > > something together for that, I can help them figure out the build system.
> > > >
> > > > The pull request is PR #10 <https://github.com/gridcf/gct/pull/10>; reviews
> > > > are welcome.
> > > >
> > > > What???s the next step?
> > > >
> > > > Thanks,
> > > > -Mat
> > > >
> > > > ???
> > > >
> > > > --
> > > > Mátyás (Mat) Selmeci
> > > > Open Science Grid Software Team / Center for High-Throughput Computing
> > > > University of Wisconsin-Madison Department of Computer Sciences
> > > >
> > > > _______________________________________________
> > > > Gt-eos mailing list
> > > > Gt-eos at mailman.egi.eu
> > > > http://mailman.egi.eu/mailman/listinfo/gt-eos
> > >
> > > --
> > > Nikhef Room H155
> > > Science Park 105 Tel. +31-20-592 5102
> > > 1098 XG Amsterdam Fax +31-20-592 5155
> > > The Netherlands Email msalle at nikhef.nl
> > > __ .. ... _._. .... ._ ... ._ ._.. ._.. .._..
> >
> >
> > > _______________________________________________
> > > Gt-eos mailing list
> > > Gt-eos at mailman.egi.eu
> > > http://mailman.egi.eu/mailman/listinfo/gt-eos
> > _______________________________________________
> > Gt-eos mailing list
> > Gt-eos at mailman.egi.eu
> > http://mailman.egi.eu/mailman/listinfo/gt-eos
>
> _______________________________________________
> Gt-eos mailing list
> Gt-eos at mailman.egi.eu
> http://mailman.egi.eu/mailman/listinfo/gt-eos
--
Nikhef Room H155
Science Park 105 Tel. +31-20-592 5102
1098 XG Amsterdam Fax +31-20-592 5155
The Netherlands Email msalle at nikhef.nl
__ .. ... _._. .... ._ ... ._ ._.. ._.. .._..
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3402 bytes
Desc: not available
URL: <http://mailman.egi.eu/pipermail/discuss/attachments/20180105/cc273984/attachment.p7s>
More information about the discuss
mailing list