git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Ramkumar Ramachandra <artagnon@gmail.com>
To: Git Mailing List <git@vger.kernel.org>
Cc: Sverre Rabbelier <srabbelier@gmail.com>,
	Jonathan Nieder <jrnieder@gmail.com>,
	David Michael Barr <david.barr@cordelta.com>,
	Daniel Shahaf <d.s@daniel.shahaf.name>,
	Sam Vilain <sam@vilain.net>, Eric Wong <normalperson@yhbt.net>,
	Junio C Hamano <gitster@pobox.com>,
	"Shawn O. Pearce" <spearce@spearce.org>
Subject: [GSoC update] git-remote-svn: Week 9
Date: Thu, 1 Jul 2010 15:13:10 +0200	[thread overview]
Message-ID: <20100701131310.GA28607@debian> (raw)

Hi,

In terms of code, this week was unproductive. However, there was one
very long and productive discussion about the design of svn-fe. We can
now prove that the design of svn-fe is most efficient, and that any
other dumpfile importer will do worse than svn-fe. I'll put this proof
down in another email, and use this one for just a status update. We
have also determined that some some extensions to git-fast-import will
significantly reduce the complexity of svn-fe. Here are the current
list of tasks along with their status:

1. The svnclient_ra. I have stopped working on a stopgap
   implementation of svnclient_ra that dumps full text/ props; to do
   this, it needs to interrupt a replay and open a new connection- it
   seems that this is quite non-trivial to get right. I've decided to
   complete the delta dump generator instead. Unfortunately, this
   means that I won't have a complete chain to show a remote helper by
   mid-term evaluations: just individual components.
2. Ternary treap fork. This only changes the underlying data
   structure- nothing else should change. David is close to completing
   this.
3. Make svn-fe support deltified dumps. This can only be done after we
   have a Git object store backing. Deferred for the moment.
4. git-fast-import only supports filecopy for trees from the parent
   commit; make it support copies from all commits. Jonathan has
   already sent an RFC patch on the list for this.
5. Zero-tree fork. svn-fe currently maintains trees for all revisions;
   Jonathan suggested that it can lazily fetch tree objects from the
   Git store backing. This change will significantly reduce the
   complexity of svn-fe.
6. git-fast-import needs to be extended to print-marks. Sverre
   suggested this, and Jonathan has sent an RFC patch on the
   list. This is required for the Git object store to communicate back
   commit hashes to the caller (svn-fe in this case).

In other news, we managed to get colabti.org to log the channel where
we have most of our discussions: #git-devel on Freenode IRC. This
information is probably helpful for other GSoC students and developers
who are looking to collaborate on projects.

-- Ram

             reply	other threads:[~2010-07-01 13:37 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-07-01 13:13 Ramkumar Ramachandra [this message]
2010-07-02  1:40 ` [GSoC update] git-remote-svn: Week 9 Bo Yang

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=20100701131310.GA28607@debian \
    --to=artagnon@gmail.com \
    --cc=d.s@daniel.shahaf.name \
    --cc=david.barr@cordelta.com \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=jrnieder@gmail.com \
    --cc=normalperson@yhbt.net \
    --cc=sam@vilain.net \
    --cc=spearce@spearce.org \
    --cc=srabbelier@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).