cabal install fails with bytestring-0.9 about hashable HOT 10 CLOSED

that seems very strange to me; using the logical negation operator is quite common in CPP conditionals

from hashable.

That seems odd to me too. Can you tell GHC to dump the preprocessed output so we can see what it looks like?

from hashable.

Probably, but please tell me how to do that...

from hashable.

would

#if !(MIN_VERSION_bytestring(0,10,0))

have worked? if so, what's the contents of the cabal_macros.h file generated by cabal? (I'm asking, because very old Cabal had fragile macros causing this very kind of problem)

from hashable.

Yes, the parenthesized version actually works! I wonder why my previous tests blamed the negation...

from hashable.

from hashable.

From my cabal_macros.h:

/* package bytestring-0.9.1.5 */
#define MIN_VERSION_bytestring(major1,major2,minor) \
  (major1) <  0 || \
  (major1) == 0 && (major2) <  9 || \
  (major1) == 0 && (major2) == 9 && (minor) <= 1

Have to transport the kids now, see you later!

from hashable.

from hashable.

Yes, it's cabal 0.8 (Debian oldstable). Upgrading cabal is also problematic. Said workaround in hashable would help me, but other necessary packages also pose problems, so I may eventually give up this GHC 6.12 environment anyway... :(

from hashable.

@tibbe btw, if you don't want to workaround this, just set cabal-version >=1.10 in future releases' hashable.cabal (that's the Cabal version that went with GHC 7.0)

from hashable.