git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Shawn O. Pearce" <spearce@spearce.org>
To: Mike Hommey <mh@glandium.org>
Cc: Sky Patrick <patricksky852@gmail.com>, git@vger.kernel.org
Subject: Re: How can i write a new protocol for git
Date: Fri, 5 Sep 2008 07:52:46 -0700	[thread overview]
Message-ID: <20080905145246.GA20537@spearce.org> (raw)
In-Reply-To: <20080905094746.GA26820@glandium.org>

Mike Hommey <mh@glandium.org> wrote:
> On Fri, Sep 05, 2008 at 05:22:57PM +0800, Sky Patrick wrote:
> > 
> > If i want to write a new protocol for file transfer and would like to
> > to integrate it with git. Is it possible and easy to integrate it ?
> > Currently, i know that jgit may achieve it, but it is a Java
> > implementation and may not work with git command line program that
> > written in C.

Writing a dumb protocol support isn't difficult in JGit due to the
abstractions in place.  Its a little harder in C Git, unless Mike
(or someone else) can abstract out the cURL library usage and make
it more flexible.  Though these days cURL supports just about any
file transfer protocol known.
 
> First, it would depend if you need this new protocol to be a dumb
> protocol (like rsync and http) or "native" (git or ssh).

Writing a smart protocol is a PITA.  Look at the smart HTTP effort
I am trying to do.  OK, I'm not spending a whole lot of time on
it, but its a damn difficult thing to do well, and to keep simple
enough that it can be described and reimplemented.  Otherwise we
have another that-thing-before-Git on our hands.

> I've been meaning to refactor the dumb protocol "stack" for a while, but
> haven't gone much further than ideas in my mind (i.e. no code). I also
> have some start of http code recfactoring, that is waiting for this dumb
> protocol rework to be completed.

Yea.  One reason why I'm doing the smart HTTP effort prototype in
JGit is because the C code structure makes this harder to implement
without first going through a number of refactorings.  Though for the
smart HTTP stuff I really need to refactor fetch-pack/send-pack and
upload-pack/receive-pack, not the dumb walkers.

-- 
Shawn.

  reply	other threads:[~2008-09-05 14:53 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-09-05  9:22 How can i write a new protocol for git Sky Patrick
2008-09-05  9:38 ` Jakub Narebski
2008-09-05  9:47 ` Mike Hommey
2008-09-05 14:52   ` Shawn O. Pearce [this message]
2008-09-08  2:32     ` Sky Patrick

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=20080905145246.GA20537@spearce.org \
    --to=spearce@spearce.org \
    --cc=git@vger.kernel.org \
    --cc=mh@glandium.org \
    --cc=patricksky852@gmail.com \
    /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;
as well as URLs for NNTP newsgroup(s).