From: David Scott <dave.scott@eu.citrix.com>
To: Anil Madhavapeddy <anil@recoil.org>
Cc: Gerd Stolpmann <info@gerd-stolpmann.de>, <r.3@libertysurf.fr>,
<caml-list@inria.fr>
Subject: Re: [Caml-list] opam and godi
Date: Mon, 22 Jul 2013 14:51:49 +0100 [thread overview]
Message-ID: <51ED38F5.2030705@eu.citrix.com> (raw)
In-Reply-To: <6796313D-9F02-4C70-87D9-8DC9BC896028@recoil.org>
On 21/07/13 15:20, Anil Madhavapeddy wrote:
[ snip ]
> Feature minutiae aside, I'd say the biggest benefit of OPAM is the more
> open development workflow. It's easier for people to maintain their
> own branches and contribute changes to the central repository.
[ snip ]
> Several development groups also maintain their own remotes without any
> need to depend on the central repository. For example, see Citrix's:
> https://github.com/xapi-project/opam-repo-dev/tree/master/packages
I really like the git-style development workflow encouraged by OPAM.
Note I'm not a GODI user so I don't know if the same thing can be done
there.
The packages you link to above are in development, and having our own
package repo in git means that we can use pull requests both to propose
changes to the code, and to quickly update the packages. As each package
matures, we can then make pull requests to the main OPAM repo.
Lowering the effort needed to create and maintain a package has had
another useful side-effect (IMHO). Previously while working on
Xen(Server) we created a few large monolithic libraries which were
tending (perhaps) towards becoming 'frameworks' (in a bad way). Now
we've started isolating distinct areas of functionality and chopping
them off into independently usable libs, or dropping them altogether in
favour of better alternatives (e.g. we're dropping a custom HTTP server
library in favour of cohttp). This kind of strong isolation between
components will make the code easier for us to maintain.
Regarding binary packages: for our use-cases binary package support via
OPAM or GODI isn't necessary. We'd like to distribute binary packages of
our software through Linux distributions, which means we're keen to
create .rpms and .debs in the usual way. Typically our stuff runs as
services, which means you need to write distro-specific code to
integrate with systemd/upstart anyway, so there's little an automatic
tool can do to help.
Cheers,
Dave Scott
XenServer System Architect, Citrix
next prev parent reply other threads:[~2013-07-22 13:51 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <51E40425.8060604@libertysurf.fr>
2013-07-15 14:20 ` [Caml-list] ocaml glade gtk3 r.3
2013-07-15 19:18 ` Martin DeMello
2013-07-19 17:22 ` [Caml-list] opam and godi r.3
2013-07-21 13:54 ` Gerd Stolpmann
2013-07-21 14:20 ` Anil Madhavapeddy
2013-07-21 14:43 ` Yaron Minsky
2013-07-22 7:44 ` Adrien Nader
2013-07-22 13:30 ` Anil Madhavapeddy
2013-07-22 8:06 ` Török Edwin
2013-07-22 12:20 ` Gerd Stolpmann
2013-07-22 12:55 ` Gerd Stolpmann
2013-07-22 18:30 ` Yaron Minsky
2013-07-22 13:51 ` David Scott [this message]
2013-07-22 6:55 ` Fabrice Le Fessant
2013-07-22 7:35 ` Francois Berenger
2013-07-22 7:46 ` Adrien Nader
2013-07-22 12:34 ` Gerd Stolpmann
2013-07-22 7:20 ` Matthieu Dubuget
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=51ED38F5.2030705@eu.citrix.com \
--to=dave.scott@eu.citrix.com \
--cc=anil@recoil.org \
--cc=caml-list@inria.fr \
--cc=info@gerd-stolpmann.de \
--cc=r.3@libertysurf.fr \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox