Re: [RFC 0/2] svn-fetch|push - an alternate approach

[Marco Schulze]

On 18-08-2012 14:39, James R. McKaskill wrote:

As a twist the code does not use the svn library, but rather talks the
svn protocol directly. I actually found it much easier to go this route
then trying to bend everything to how the svn library understands
things. It also has the advantage of not depending on libsvn. A number
of distributions currently distribute the svn specific parts of git
seperately to avoid this dependency.
It is a bit of a pain to work with libsvn, but I think it is worth using 
it, not only due to protocol support, but also due to ready-to-use 
support for SVN deltas and dumps.  Pipelining could be implemented by 
maintaining a set of connections (svn_ra_session_t structures) and 
manually distributing work between them.

--
To unsubscribe from this list: send the line unsubscribe git in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Issuing cat-blob or ls to fast-import from a remote helper

[Marco Schulze]

I've been toying with a git-remote-svn which uses fast-import (by 
advertising the 'import' capability) to get data into Git. 
Unfortunately, I can't get the result of any commands issued to 
fast-import, as its stdout is not redirected, and the remote helper API 
offers no alternative.


I tried modifiying transport-helper.c to force redirection 
(fast-import's stdout - remote helper's stdin), but that breaks test 
t5800 (and possibly others). Is this a bug, or intended behaviour? Am I 
missing something?

--
To unsubscribe from this list: send the line unsubscribe git in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html