Re: [RFC/PATCH] drop vcs-svn experiment

From: Jonathan Nieder <jrnieder@gmail.com>
Date: Aug 18, 2018, 05:26 UTC
Message-ID: <20180818052605.GA241538@aiede.svl.corp.google.com>
In-Reply-To: <20180817190310.GA5360@sigill.intra.peff.net>

Hi,

Jeff King wrote:

Show 5 quoted lines

> The code in vcs-svn was started in 2010 as an attempt to
> build a remote-helper for interacting with svn repositories
> (as opposed to git-svn). However, we never got as far as
> shipping a mature remote helper, and the last substantive
> commit was e99d012a6bc in 2012.

I do use svn-fe occasionally, and have done so in the past few years. That said, it's probably not worth keeping this in tree just for me.

> We do have a git-remote-testsvn, and it is even installed as
> part of "make install".

At a minimum, we should stop doing that.

[...]

> We also ship contrib/svn-fe, which builds on the vcs-svn
> work. However, it does not seem to build out of the box for
> me, as the link step misses some required libraries for
> using libgit.a.

What libraries do you mean?  It builds and runs fine for me with

 $ git diff
 diff --git i/contrib/svn-fe/Makefile w/contrib/svn-fe/Makefile
 index e8651aaf4b5..bd709f8d83b 100644
 --- i/contrib/svn-fe/Makefile
 +++ w/contrib/svn-fe/Makefile
 @@ -4,7 +4,7 @@ CC = cc
  RM = rm -f
  MV = mv
 
 -CFLAGS = -g -O2 -Wall
 +CFLAGS = -g -O2 -Wall -pthread
  LDFLAGS =
  EXTLIBS = -lz

which appears to be platform related, not due to some internal change in Git.

[...]

> Of course, I could be completely wrong about people using this. Maybe
> svn-fe builds are just completely broken on my system, and maybe people
> really do use testsvn::. But if so, they certainly aren't talking about
> it on the mailing list. :)

My take:

 - svn-fe works fine and has been useful to me, though its Makefile
   could likely be simplified and made more user-friendly

 - I've benefited from the test coverage of having this in-tree

 - testsvn:: is a demo and at a minimum we ought not to install it
   with "make install"

 - keeping this in-tree for the benefit of just one user is excessive,
   so removing it is probably the right thing

 - it would be nice if the commit removing this code from Git includes
   a note to help people find its new home

Would you mind holding off until I'm able to arrange that last bit?

Thanks, Jonathan