Re: [PATCH 0/6] Another attempt to get the SVN exporter merged

From: Jonathan Nieder <jrnieder@gmail.com>
Date: Jun 12, 2010, 06:26 UTC
Message-ID: <20100612062613.GA2549@burratino>
In-Reply-To: <1276175389-6185-1-git-send-email-artagnon@gmail.com>

Hi Ram,

Ramkumar Ramachandra wrote:

> It certainly looks like I'll never give up trying to get this series
> merged- this is my third attempt

First of all, thanks for your work on this and sorry to take so long to respond[1].

> I still haven't been able to get rid
> of the compiler warnings about unused functions

Will reply to the relevant patch. Since it’s not so bad to fix, I don’t think that should hold up queueing the series in pu.

> it looks like
> this series won't graduate to `master` before that happens- I'd
> appreciate pointers on how to do this.

Instead, I suspect the main missing ingredient is a caller. The series builds up some library infrastructure with no new feature to exercise it, which makes it a good time to get feedback but not a good time to merge.

Of course, we could easily add a feature to exercise it: David’s svn-fe tool is IMO quite useful on its own (regardless of how the code will be used later by remote-svn).

I don’t think that is so important for the remote-svn project. In other words, I hope some other reviewers show up, but if that doesn’t happen, I would suggest submitting again for inclusion once the remote-svn command with import capability is functional.

I planned to write a series adding svn-fe to contrib/ and then the computer died. If someone else doesn’t do it first, hopefully I can try again this weekend.

> Major change since last time: Removed dependency on mmap for
> portability reasons.

Sad but perhaps necessary. I guess this also opens the possibility of later making the save operation atomic with the fsync() + rename() trick.

Regards, Jonathan

[1] My excuse: laptop died. My new setup is finally in place and working well but this Monday I will be on vacation in a land of poor connectivity --- agh.