logo       

build of GHCi 5.04.3 fails to work (libgcc.a): msg#00081

lang.haskell.glasgow.bugs

Subject: build of GHCi 5.04.3 fails to work (libgcc.a)

Downloaded, unpacked GHC release 5.04.3,
./configure
echo "GhcWithInterpreter=YES" > mk/build.mk
make
make install

ghci
___ ___ _
/ _ \ /\ /\/ __(_)
/ /_\// /_/ / / | | GHC Interactive, version 5.04.3, for Haskell 98.
/ /_\\/ __ / /___| | http://www.haskell.org/ghc/
\____/\/ /_/\____/|_| Type :? for help.

Loading package base ... linking ... /usr/local/lib/ghc-5.04.3/HSbase_cbits.o:
unknown symbol `__umoddi3'
ghc-5.04.3: panic! (the `impossible' happened, GHC version 5.04.3):
can't load package `base'

Please report it as a compiler bug to glasgow-haskell-bugs@xxxxxxxxxxx,
or http://sourceforge.net/projects/ghc/.

which looks like HSbase_cbits.o isn't getting linked with libgcc.a for
some reason.

fixing this with
mv HSbase_cbits.o HSbase_cbits.orig.o
ld -r -o HSbase_cbits.o HSbase_cbits.orig.o
/usr/lib/gcc-lib/i386-linux/3.2.3/libgcc.a

lets everything work fine, though when I tried on an earlier build, it
didn't. (Instead it complained about not being able to load DLLs, I
think. I really should have written it down.)

I am using debian gcc-3.2.3, which hasn't had as much pounding as it
should have, and I did add "SRC_HC_OPTS += -ldl" to mk/build.mk as well,
as I noticed in mk/config.mk that the GHCi enabling section did this.
It only seemed to end up with extra -ldl flags though.

I see on the mailing list reports of other problems that look like
libgcc.a not getting linked in. Here's another datapoint.

--
Aaron Denney
-><-


<Prev in Thread] Current Thread [Next in Thread>
Google Custom Search

News | FAQ | advertise