Running "emerge ghc" on my up-to-date Gentoo 1.4rc2 based linux system (compiled from source) fails to compile and install the Glasgow Haskell Compiler. -------------------------------------- It fails during the compilation phase, issuing a large number of messages like: CTypes.hc:48547: warning: excess elements in array initializer CTypes.hc:48547: warning: (near initialization for `CTypes_tpl767_closure.payload') ... and finishing with: CTypes.hc:55324: warning: excess elements in array initializer CTypes.hc:55324: warning: (near initialization for `CTypes_zdfReadCChar_closure.payload') Prologue junk?: czvw_ret: movl $0, 16(%esp) movl $0, 20(%esp) movl $0, 8(%esp) movl $0, 12(%esp) make[2]: *** [CTypes.o] Error 25 make[1]: *** [all] Error 1 make: *** [all] Error 1 !!! ERROR: dev-lang/ghc-5.04 failed. !!! Function build_stage1, Line 228, Exitcode 1 !!! (no error message) ------------------------------- Looking through the messages the ebuild produces at the beginning, I wonder whether this has to do with the fact that my machine has no pre-existing GHC compiler installed, so that the ebuild has to do some sort of bootstrap; if the ebuild is tested on a machine that already has GHC, it may follow a different path and perhaps succeed.
Can't get around to this anytime soon, sorry.
Hi Phil. Do you still experience this problem? The ghc has been fixed since, 5.04.3 is a latest version, however if you are running the latest glibc-2.3.2, which is still in testing profile, you might want to stick to ghc-5.04.2. The ghci of 5.04.3 version has some problems dynamically linking against newer glibc on some systems (package builds fine and the compiler works, only ghci is problemmatic). Well, this problem is spotty, so there are chances of ghci in 5.04.2 having the same misbehavior. Hopefully we will get a fix soon. And you can always use ghc-bin. Andres: This bug should really belong to you now ;). Mind reassingning it to yourself? Also, unfortunately I did not get to testing your fix yet, will do so today and let you know.. George
Hi Andres. I have tested your patch - works wonders! Some output below. George [gerr@groug gerr]$ ghci ___ ___ _ / _ \ /\ /\/ __(_) / /_\// /_/ / / | | GHC Interactive, version 5.04.3, for Haskell 98. / /_\\/ __ / /___| | http://www.haskell.org/ghc/ \____/\/ /_/\____/|_| Type :? for help. Loading package base ... linking ... done. Loading package haskell98 ... linking ... done. Prelude> Leaving GHCi. [gerr@groug gerr]$ [gerr@groug gerr]$ ldd "/usr/lib/ghc-5.04.3/ghc-5.04.3" libdl.so.2 => /lib/libdl.so.2 (0x4002b000) libreadline.so.4 => /lib/libreadline.so.4 (0x4002e000) libncurses.so.5 => /lib/libncurses.so.5 (0x4005a000) libm.so.6 => /lib/libm.so.6 (0x4009c000) libgmp.so.3 => /usr/lib/libgmp.so.3 (0x400be000) libc.so.6 => /lib/libc.so.6 (0x400ed000) /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000)
I am closing this bug. I assume that it works with the newer ebuilds for GHC.
so close it really ;)