[linux-elitists] Contributing to a Project with a Maintainer Who Doesn't Merge Contributions Quickly

Shlomi Fish shlomif at iglu.org.il
Mon Mar 9 11:55:21 PDT 2009


Sorry for the late response.

On Saturday 17 January 2009 20:22:38 Don Marti wrote:
> begin Shlomi Fish quotation of Sat, Jan 17, 2009 at 07:02:14PM +0200:
> > So now I'll have to study the python setuptools just to fix the bzr-svn
> > build.
> There's always git, which is a DVCS with fewer
> dependencies, mostly in portable C.  (Now a
> mix of C and shell, with a little Perl, they're
> rewriting more and more of it in C every release.
> Other dependencies: zlib, openssl, libcurl, ssh,
> expat; plus Tcl/Tk if you want gitk and git-gui)

Well, it's not as if bzr-svn has so much more dependencies, that I am unable 
to resolve with urpmi in Mandriva. It's just that its rpm creation is borked, 
which is easy to fix given the good know-how. If a software package does not 
work, the proper action for a capable hacker to do is to see how to remedy it. 
Like I say:

Hacker sees bug.
Hacker does not want bug.
Hacker fixes bug.

I already fixed the rpm creation of subvertpy which is included in bzr-svn:


And I now intend to take a look at bzr-svn itself. Then I can choose between 
bzr, git and other candidates that can work against a Subversion repository.

All that put aside, I may be forced to resort to git to contribute to gdb, 
which is sadly still using CVS "due to inertia". Last time I tried to git-
cvsimport its CVS repository from the most recent month at a time, the process 
hanged-up in the middle. :-(

> > In any case, I have some philosophical and hypothetical reservations on
> > using a private repository like that.
> You don't have to think hypothetically about this
> stuff -- you could find one of the projects using a
> DVCS (list for git here:
>   http://git.or.cz/gitwiki/GitProjects
> ) and lurk on the mailing list.  It seems like the
> problems you describe come up often, and DVCS users
> are using a variety of approaches, technical and
> social, to handle them.

Well, even though I'm unemployed at the moment, I have better things to do 
with my time, than lurk on mailing lists of projects I have no interest in, 
just so I can better work on one project that's not too important to me to 
begin with. If you can explain them here and now, please do.

> > Furthermore, if I keep a repository full of my own private changes that
> > were not integrated yet, and that I plan to submit one after the other,
> > then it's essentially a fork, and I'd like to avoid that.
> But you get to draw a groovy graph of all the
> micro-forks and re-merges!


> The original post in this thread got two independent
> replies, from Asheesh Laroia and Ben Finney, both
> suggesting going DVCS and letting the "maintainer"
> catch up if he or she wants to.  I know there a lot
> of projects that are starting with or moving to DVCS
> culture, but anyone know of an example in the wild
> of DVCS sedition in an existing CVS/SVN-based project?

Let me continue with the story. I decided to confront the project owner and 
ask for giving me a commit bit for the project's repository (on IM). He told 
me that he'd rather that I send him patches. Then I promised him I'll only 
work on a separate branch or branches, and told him I have a problem with the 
long round-trip delay. Then I asked him if he can think of any other solution. 
He said he doesn't have any need to think of a different solution, and that I 
should just send him patches.

After I told all this to a different person, he told me that I should use git 
or SVK to interact with the remote Subversion repository without a commit bit. 
Since several people already recommended this solution, I guess that settles 
it and that's what I'll do.

I don't know why the owner of the project in question exhibits this attitude, 
but I think that, given it, he just *wants* me to fork it. I'm not going to, 
because I'd like to give the DVCS solution a try, and because that project is 
not so critical to me, but under different circumstances, I may have actually 
gone the full monty and forked it.


	Shlomi Fish

P.S: Brand new quote in the signature. I liked my previous signature (with the 
IRC conversation) a lot but it was kinda long.

Shlomi Fish       http://www.shlomifish.org/
"The Human Hacking Field Guide" - http://xrl.us/bjn8q

God gave us two eyes and ten fingers so we will type five times as much as we

More information about the linux-elitists mailing list