https://bugzilla.redhat.com/show_bug.cgi?id=1273579
--- Comment #14 from Zbigniew Jędrzejewski-Szmek <zbyszek(a)in.waw.pl> ---
(In reply to Ankur Sinha (FranciscoD) from comment #13)
(In reply to Zbigniew Jędrzejewski-Szmek from comment #12)
> (In reply to Ankur Sinha (FranciscoD) from comment #11)
> > I'm happy to continue the review if another reviewer comes along,
> I don't understand this sentence. Raphael said that he can review the
> package in comment #8. If he cannot do it, I'm sure that somebody else (e.g.
> me) would step up.
Oh, he dropped the review after I asked his opinion about the build system
change in comment 9 which I took as an indication that he agreed with me
putting off the package (He isn't the assignee any more, and he isn't in the
CC list either) :)
Was he ever? I only see comment #8 from him.
Anyway, like I said, I can review the package.
> > but I do
> > think a re-review will be in order (even if unofficially) if the build
> > system and change.
> Nah, the build system is not visible in the binary package. As long as it
> builds nobody cares (except the maintainer of course).
That is what I meant - when a new build system comes, the entire package has
to be effectively redone, and nest, as you'll see from the spec, isn't the
simplest of packages. Totally a maintainers burden, yes, but I'm going to be
the maintainer ;)
True. I doesn't seem so bad to me (about normal for an mpi
package in fact :)),
but of course I can't and don't want to push you.
> > Is the plan to get all the neuroscience packages into
Fedora by F24, though
> > - or can we have a "group copr" thing running and transition packages
over
> > as they pass review?
>
> I don't think there's a plan to get "all" neuroscience packages
into Fedora,
> there's probably too many to even consider that.
>
https://fedoraproject.org/wiki/Changes/NeuroFedora doesn't have a specific
> list. New packages can be added at any time, and there's quite a bit of time
> before Fedora 24 release (2016-05-17). I don't know about any plans to have
> a group copr. I see coprs as a good mechanism to provide alternative
> versions of already packaged software, or often updated software, or things
> which are inappropriate for main Fedora for other reasons. I don't think
> nest or other science related software falls into any of those categories,
> and making a "detour" through copr would be mostly a waste of time.
I think the list is somewhere on a google spreadsheet. I don't have the link
handy at the moment.
I wasn't aware of that. I know there's
http://taiga.fedorainfracloud.org/project/ignatenkobrain-neurofedora/kanban
but probably not everything is on there.
Coprs are also a good mechanism for packages that build and are
functional
but are not packaged well enough to pass a formal review yet (outer ring in
fedora.next and all that), which is exactly what the nest package is at the
moment.
In my experience, it's quite easy for coprs to get obsolete. If the
package
is in Fedora proper, than at least it will be regularly rebuilt, sometimes
a provenpackager will do a drive-by fix of some small issue, and users can
report issues in bugzilla, crashes are colleted. I think coprs are great
for development, or for a small focuses group of users, but after using them
for a while, I'm starting to see downsides of using them as a mechanism for
distributing software to a diverse group of users, over long time.
Anyway, I'm happy to continue the review if someone takes it up.
I'll update
the spec with Raphael's comments later this week. I already updated the copr
package to 2.10 which came out recently. I need to work on the python3 bit
etc.
Great! Let's see if Raphael wants to pick up the review. If not, I'll take it.
--
You are receiving this mail because:
You are on the CC list for the bug.
You are always notified about changes to this product and component