1) If Cabal find the dependencies aren't satisfied then it puts
  "buildable: False" somewhere and exits successfully. I quite like
  this from a consistency point of view (it's what happens when
  ./configure decides that the package is unbuildable).

this is also what Cabal setup already does, eg, for the OpenAL binding, when no OpenAL installation is found, so i don't understand why the setup behaviour
differs when ALUT cannot find an OpenAL binding?

i couldn't find any hook to tune what setup configure does on failure, the call
to die seems hardcoded. using confHook to wrap configure and catch the ExitFailure exception seems like a hack, but got me somewhat further:

   $ darcs whatsnew
   plink: unknown option "-O"
   {
   hunk ./Setup.hs 6
   +import Distribution.Simple.Configure(configure)
   hunk ./Setup.hs 11
   +import qualified Control.Exception as CE (catch)
   +import System.Exit
   hunk ./Setup.hs 19
   +                  confHook = \pd cf->CE.catch (configure pd cf) noBuild,
   hunk ./Setup.hs 23
   +          where noBuild e = do writeHookedBuildInfo "ALUT.buildinfo" $
   +                                   (Just emptyBuildInfo{buildable=False},[])
   +                               exitWith (ExitSuccess)
   +                               return undefined
   }

now, make falls over in setup haddock. noBuild could write a dummy .setup-config, i guess, but perhap setup haddock should be wrapped in ifBuildable instead?

claus

---------------------
   cd OpenAL && setup/Setup haddock
   Reading parameters from c:\fptools\ghc\libraries\OpenAL\OpenAL.buildinfo
Running Haddock for OpenAL-1.3... cd ALUT && setup/Setup haddock
   Reading parameters from c:\fptools\ghc\libraries\ALUT\ALUT.buildinfo
Setup.exe: error reading ./.setup-config; run "setup configure" command? make[1]: *** [doc.library.ALUT] Error 1
   make[1]: Leaving directory `/cygdrive/c/fptools/ghc/libraries'
   make: *** [stage1] Error 2

_______________________________________________
Cvs-ghc mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/cvs-ghc

Reply via email to