id,summary,reporter,owner,description,type,status,priority,milestone,component,version,resolution,keywords,cc,os,architecture,failure,difficulty,testcase,blockedby,blocking,related
4914,FPU initialization required again,aruiz,simonmar,"I think that the bug [http://hackage.haskell.org/trac/ghc/ticket/2724 2724], fixed in ghc-6.10.x, has reappeared in ghc-7.0.1. When I run hmatrix tests I get the same kind of random errors, which disappear if I insert the asm finit call. I don't have a small test case without dependencies, but I hope that this problem can be easily reproduced as follows:


{{{
$ cabal update
$ cabal unpack hmatrix
$ cd hmatrix-0.11.0.0
$ cabal install [-f-vector] [-f-binary]
}}}


(these flags can be given to minimize dependencies)


{{{
$ cabal test
}}}


All tests must pass. Now we will try to produce the error:

Comment out line 172 of hmatrix.cabal to remove the workaround for ghc-7.0.1:


{{{
-       cpp-options: -DFINIT
+    -- cpp-options: -DFINIT
}}}



{{{
$ cabal clean
$ cabal install [-f-vector] [-f-binary]
$ cabal test
}}}


Now we will probably get several errors.

Please let me know if you need more information. I hope I have not made some big mistake.
",bug,closed,high,7.2.1,Compiler (NCG),7.0.1,fixed,,fryguybob@…,Unknown/Multiple,x86,Incorrect result at runtime,,,,,