[darcs-users] wishlist for darcs 2.1

[Eric Kow]

Hi everyone,

It may be time to start thinking about the next version of darcs (say,
for early 2009).

I thought I might put down my personal wishlist and see what you
think. My goal here is to have a very small number of well-defined
objectives.  Please keep in mind that, as usual, I am speaking as a
regular user, and not in any official capacity whatsoever.

(1) critical/urgent bugs - darcs 2.0.0 seems much more robust than
darcs 1, but we still do have at least one case where it blows up
(issue857) and another where doing funny things can have surprising
results (issue687).

(2) performance issues - Aside from fixing the critical issues, it
would be nice if we focused our energy on performance issues.  As I
understand things, we have had deep problems (patch theory related)
entangled with day to day problems.  Now that the deep stuff has been
taken care of (less chance of conflict misery), we should focus on the
day-to-day.
Looking at the Performance topic in the bug tracker suggests that our
main issue is that darcs uses/leaks a lot of memory.

Maybe we can have a concrete release goal, like using only half as
much memory on memory-related test cases by darcs 2.1.

Speaking of which, we still need some sort of memory suite, even one
that tells us our usage for some common test cases.
  http://bugs.darcs.net/issue99
Something simple like this would be a really great first step!

(3) A third item I would like to see on our agenda are improvements to
the Windows-darcs experience.  Darcs still has some annoying
behaviours on Windows, for example, not echoing your choices in
interactive prompts, or not noticing when programs, like Emacs are not
found.  These should be easy to fix, except that none of us are
regular Windows users...

If you want to help darcs, I would say the best way you can make an
impact is (a) to contribute a memory-suite and tests for it and (b) to
propose your Windows/Haskell hacking skills (if you don't know
Haskell, but you do know Windows, I'm sure that would still be
useful).

Those are my personal desires anyway.
What do you think?  Sound like a plan?

-- 
Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow>
PGP Key ID: 08AC04F9