[opam-devel] [Caml-list] [ANN] IOCaml v0.3

[Anil Madhavapeddy]

On 14 Feb 2014, at 15:27, Sylvain Le Gall <sylvain+ocaml at le-gall.net> wrote:
> 
> Well, it is one of the thing I don't like about github:
> 
> https://github.com/pdhborges/ocaml-zmq
> 
> And everything seems to be forked from this repository... So it looks
> to me that the official repository is this one and that the owner is
> pdhborges! You can argue that you can look at the network graph to see
> that another one is more active and in advance, but this network graph
> is super long to load and when it works, it is not that obvious who is
> in advance....
> 
> But maybe I am missing the point here.
> 
> I don't like having my name in the URL for my project, because I don't
> feel I have a strong ownership of it (or at least I would like to
> invite anyone to participate).

Note that it's possible to just mail GitHub support to disconnect your
repo from the fork graph.  I've done this for several Mirage libraries
that had to be switched around.

As for ownership, we just set up an organization when such things are
required (again, see http://github.com/mirage or many others for how
this works).

>> Like Anil, I'm not exactly sure how forge helps in this case as Active
>> development would still occur on github. And I guess the idea is to allow
>> users to understand where to find most resent sources. Opam should be able
>> to provide a reference to the relevant fork on github.
>> 
> 
> I was suggesting the forge just for hosting tarballs, no dev involved there.
> 
> And indeed, OPAM can point to the official repository. But whenever
> someone will look for ocaml-zmq on Google the first link will be
> pdhborges/ocaml-zmq for at least a couple of months and most people
> may assume that it is the official one. If you update this in OPAM,
> you will have to go to OPAM read the page and realize that the
> official one is not the one that Google reports.... I just think this
> is misleading. But this is just about communication, no dev here.

This is indeed a good point, which is why I raised the question of
transition here rather than just commit the change.  The opam.ocaml.org
web-ranking is improving though, so it should be reasonably obvious
that running an "opam search zmq" will be the latest version.  We
shouldn't leave two concurrent versions of the library in opam stable
for too long.

-anil