Am 14.02.2015 um 17:02 schrieb Jens Petersen:
AFAICT now the ABI changes are due to -40 having
been built with make -j16 on i686 and x86_64.
In future I will use -j4 or less again
don't get me wrong but if the number of threads of the complier changes
ABI of the result the problem is
ghc-7.8.4-41.fc22
-
Oh dear - gcc5 changes all the ghc library ABI hashes.
Sorry my initial hunch seems incorrect.
AFAICT now the ABI changes are due to -40 having
been built with make -j16 on i686 and x86_64.
In future I will use -j4 or less again.
Jens
--
devel mailing
:
[alex]
alex-3.1.3-1.fc22.i686 requires
ghc(unix-2.7.0.1-b0d741fe9ce9a85eada7f5178fc621fa)
alex-3.1.3-1.fc22.i686 requires
ghc(time-1.4.2-c7bfcef03445b2ec67d3d5b04c386722)
alex-3.1.3-1.fc22.i686 requires
ghc(template-haskell-2.9.0.0-57f65c241f99343438be3dddb7391322)