[Haskell-cafe] poll: how can we help you contribute to darcs?

Claus Reinke claus.reinke at talk21.com
Mon Aug 4 09:37:37 EDT 2008

> It would be useful if some darcs 2 hackers, contributors could help the
> ghc people evaluate if darcs 2 is still in the running. That would mean
> identifying the key bugs (eg windows case-insensitive file bugs, slow
> pulls) and seeing how hard they are to fix. Also doing a test conversion
> to darcs 2 format and benchmarking some of the key operations described
> on the ghc evaluation page.
> http://hackage.haskell.org/trac/ghc/wiki/DarcsEvaluation

I think Simon Marlow has been running a converted ghc repo 
for a while, reporting performance and other issues arising;
he even put up a fixed darcs 2 binary for windows.

One thing I have been wondering about, assuming that darcs 2
(with darcs 2 format) really fixes the bugs that make darcs 1 such 
a headache with ghc, so that performance on large repos would 
be the major remaining hurdle, is this:

- according to specs, all darcs repos are equal, patches have 
    no particular order, other than calculated dependencies, and
    versions are just sets of patches

- in practice, major projects tend to have a reference repo,
    and lots of developer or branch repos; in the ghc case, any
    successful patch ultimately makes it into the main repo and
    is pulled from there into every other repo (some of which
    might have the patch already)

Could darcs 2 performance be improved by making use of the
order of patches in the reference repo, to identify reference
versions and reign in exponential permutation issues? In other
words, all repos are equal, all patches are equal, but once a patch has made the
roundtrip through the reference repo, it i


More information about the Haskell-Cafe mailing list