On 6/14/07, Neil Mitchell <[EMAIL PROTECTED]> wrote:
Hi Creighton, > This might be a silly question for you guys, but is there a reason why yhc > still uses PackedString in places instead of ByteString? Two reasons: 1) ByteStrings are totally unreliable - they can't be made to work in GHC 6.4 and GHC 6.6 without features of Cabal that aren't even written yet. We did move to Data.Binary, which requires ByteStrings, but had to change back because it was a nightmare. If the benefits were compelling we could move, but it wasn't worth it just for binary serialisation. 2) When the code was written, there weren't ByteStrings, and no one has had the time to port to them. If someone ported Yhc to ByteStrings, we'd definitely accept a patch!
Well, if you don't mind I'd like to do it. I'm trying to understand your sourcecode anyway, so I might as well do something useful while I'm fussing with it. I guess I didn't really know about these reliability issues with ByteString though, do you have some anecdotes or references about it? Cheers
_______________________________________________ Yhc mailing list [email protected] http://www.haskell.org/mailman/listinfo/yhc
