syntaxfree: replace the second 'x' by a zer0
syntaxfree: I think. Or remove it.
|dcoutts_||Cale, it must be happening because Cabal is guessing wrongly about how many args it can fit into the space allowed. Cabal already has logic to call ld and ar multiple times when there are 1000's of little .o files.|
|Cale||What's the correct way to do it?|
Do you know what the magic number mismatch thing means?
I just tried doing ar q libname.a <first half of .o's>; ar q libname.a <second half>
|Saizan||the 7.1 tarball compiled fine on win32+msys btw|
(ah ghc has it's own ar..)
|dcoutts_||Cale, yep, calling ar on the same .a file is the right thing to do (btw, 'q' does nothing|
|Cale||dcoutts_: okay, then the problem was something else.|
Bad interface file: /usr/local/lib/hxt-7.1/Text/XML/HXT/Arrow.hi
magic number mismatch: old/corrupt interface file?
Oh, great, now it's happening with 7.0 as well.
Perhaps this is a GHC 6.6 thing?
|dcoutts_||Cale, that's nothing to do with ar|
|Cale||okay, I was misled by that being the point in the install where things failed.|
What does that message actually mean?
|dcoutts_||it means ghc can't read the .hi file|
it's trying to compile something that uses that module and so reads the .hi file
|action||dcoutts_ -> off to teach|
|dcoutts_||-> off to teach|
|Eighty||is there an assert function in the standard libraries?|
|lambdabot||Control.Exception.assert :: Bool -> a -> a|
Test.HUnit.Base.assert :: Assertable t => t -> Assertion
Control.Exception.assertions :: Exception -> Maybe String
|ndm||Eighty: yes, the first one :)|
|syntaxfree||I don't understand wht Parsec doesn't seem to have a simple "negate" parser combinator.|
succeed if the parser argument fails, return empty string.
Cale: great comment!
I'll respond in place for posterity.
|twanvl||syntaxfree: negate a = try (a >> fail) `mplus` return () ?|
11:15 < syntaxfree> Cale: great comment!
11:15 < syntaxfree> I'll respo)] [Act: 1,2]] [Act: 1,2]