[darcs-users] darcs patch: Use index-based diffing in Record. (and 74 more)

Eric Kow kowey at darcs.net
Wed Oct 7 12:33:29 UTC 2009


On Tue, Oct 06, 2009 at 23:08:07 +0200, Petr Rockai wrote:
> this is a bundle of darcs-hs versus current darcs.net mainline (modulo the
> cabal rename patch). I have just uploaded hashed-storage 0.4 to hackage, so I
> think the light is green -- fire at will. (Don't forget to go through the
> conflict resolutions before pushing though. Also, sanity checking the whole
> bundle is recommended.)

I've had a cursory look in a graphical diff, and I did not notice any
functions called 'killKittens'.

I'm almost ready to push this (thanks to Ganesh's go-ahead from his
previous message), but there is one issue from the GHC-6.8 staging
I dusted off for this purpose

$ cabal install -fdeps-only
Resolving dependencies...
cabal: cannot configure dataenc-0.13.0.0. It requires base >=4.0.0 && <4.2
For the dependency on base >=4.0.0 && <4.2 there are these packages:
base-4.0.0.0 and base-4.1.0.0. However none of them are available.
base-4.0.0.0 was excluded because base-3.0.2.0 was selected instead
base-4.0.0.0 was excluded because darcs-2.3.1 requires base ==3.* && >=3 &&
==3.*
base-4.0.0.0 was excluded because of the top level dependency base -any
base-4.1.0.0 was excluded because base-3.0.2.0 was selected instead
base-4.1.0.0 was excluded because darcs-2.3.1 requires base ==3.* && >=3 &&
==3.*
base-4.1.0.0 was excluded because of the top level dependency base -any

Is this easy to rectify?  If not, are we prepared to force GHC 6.10 on anybody
who wants to build Darcs?

Also... (below is a non-blocker in my opinion)

Get rid of wait_a_moment, since we do not use pristine timestamps anymore.
--------------------------------------------------------------------------
> -wait_a_moment :: IO ()
> -wait_a_moment = do { sleep 1; return () }
> -    -- HACKERY: In ghc 6.1, sleep has the type signature IO Int; it
> -    -- returns an integer just like sleep(3) does. To stay compatible
> -    -- with older versions, though, we just ignore sleep's return
> -    -- value. Hackery, like I said.

I didn't see this mentioned anywhere, but I trust the two of you have
worked out how this impacts old-fashioned repositories and whether or
not we actually need to care...

-- 
Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow>
PGP Key ID: 08AC04F9
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: Digital signature
URL: <http://lists.osuosl.org/pipermail/darcs-users/attachments/20091007/75a118a6/attachment.pgp>


More information about the darcs-users mailing list