On Wed, 2007-10-17 at 23:37 +0100, Ian Lynagh wrote: > On Wed, Oct 17, 2007 at 06:48:57PM +0100, Duncan Coutts wrote: > > > > code.haskell.org/bytestring has a MOTD telling people to ignore it and > > use darcs.haskell.org/bytestring instead. > > Why do you prefer d.h.o over c.h.o, out of interest? > > c.h.o has the advantage that it's much easier to add accounts for > people who might want to contribute in the future.
I really don't care. But whichever we pick, we should do the same for different packages in the same category. This is important so we know which branch is the HEAD branch and so we don't get insanely confused. Currently we're being fairly consistent with the core packages. As far as I understand it, for most of the core packages and indeed many of the extralibs the HEAD branch is at d.h.o/packages/$pkg Then for bytestring and Cabal, they have their HEAD branch at d.h.o/$pkg and a stable ghc-HEAD copy that lags behind at d.h.o/packages/$pkg All of them have their stable ghc-6.8 branch at d.h.o/ghc-6.8/packages/$pkg Is that correct? We should add a rcs field to the .cabal file just so we can remember where the HEAD branches are. Especially for the extralibs packages that I was breaking^Hfixing today, it's not always clear which is the HEAD branch, or if it's even on darcs.haskell.org at all. > If you do stick with d.h.o then the c.h.o repo is sufficiently new and > (as far as I know) unadvertised that it would be better just to remove > it IMO. True. Duncan _______________________________________________ Cvs-ghc mailing list [email protected] http://www.haskell.org/mailman/listinfo/cvs-ghc
