[darcs-users] unique features + exponential time issue

[Alexander Staubo]

On 10/18/07, Stephen J. Turnbull <stephen at xemacs.org> wrote:
> > [1] All the fancy talk about patch theory comes to nought when you get
> > bogus conflicts, like when *nothing was changed*. Or when you inserted
> > code into the same spot as someone else. It happens all the time here.
>
> This is a FAQ, answered in the manual.  This can be a genuine bug, and
> since Darcs only has text-manipulation (rather than program semantics
> manipulation) patches, it can't tell the difference.  David chose the
> conservative route.  You're welcome to disagree with his choice here
> (I do, myself), but it follows directly from patch theory.

Yep, but you missed my point. There are technical reasons for all
these glaring flaws, in the same way there are technical reasons for
the existence of bugs. But they're *flaws*. We are humans, not patch
operators.

For this discussion I am not really terribly interested in the whys or
the whats -- why nobody is doing anything with these problems, or why
nobody seems to consider them important. Nor am I interested in
whether anyone else cares about the same issues as I do.

The original poster's question's was, "what advantages can darcs claim
over mercurial and git?" Since others have pointed out the good stuff,
I feel that it's pertinent to point out some of the disadvantages,
too. To wit:

(1) I consider Darcs to suffer from several showstopper bugs, some of
which I have enumerated.

(2) I think there's little interest and/or impetus in solving any of
them. (I know David Roundy has been putting in some work on the
conflict misery problem, though -- kudos!)

(3) I consider Darcs to be nearly dead in terms of major feature
development right now. (Please correct me if I'm wrong on this; in the
absence of an official development roadmap, I'm relying on what I
glean from changelogs and ML archives.)

These are factors to consider when adopting a new version control
system. I would warn any newcomer to these issues in the interest of
full disclosure.

> > [2] Darcs stalls almost consistently on Darwin/Intel when using
> > OpenSSH's control master support. This is presumably the reason why
> > it's off by default now.
>
> That will have to be fixed by someone working on that platform.  Why not you?

Easy -- this one's way over my head.

The bug seems to relate to something deep in GHC's interface with the
*BSD core libraries. I believe Eric Kow spent a lot of time on this
problem, and according to the changelog he was the one who eventually
disabled the flag ("it seems to hang on large repositories" -- so
perhaps he gave up). If a seasoned Haskell developer can't fix the
issue, then I'm not even going to consider thinking about possibly
trying.

(Bugs that nobody understands are scary, though.)

Alexander.