[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Gcl-devel] GCL internal error (GCC error on generated C code)

From: Robert Dodier
Subject: [Gcl-devel] GCL internal error (GCC error on generated C code)
Date: Sat, 3 Feb 2007 16:56:47 -0700


I'm trying to compile the current cvs version of Maxima.
It builds OK and passes its test suit with Clisp and SBCL.
GCL 2.6.7 fails on suprv1.lisp
with the following message:

Compiling /home/robert/tmp/maxima-head/maxima/src/suprv1.lisp.
End of Pass 1.
End of Pass 2.
binary-gcl/suprv1.c: In function `LC110':
binary-gcl/suprv1.c:10674: error: `Vstandard_output' undeclared (first
use in this function)
binary-gcl/suprv1.c:10674: error: (Each undeclared identifier is
reported only once
binary-gcl/suprv1.c:10674: error: for each function it appears in.)
(SYSTEM "gcc -c -Wall -DVOL=volatile -fsigned-char -pipe
-I/usr/local/lib/gcl-2.6.7/unixport/../h  -O3 -fomit-frame-pointer -c
\"binary-gcl/suprv1.c\" -o \"binary-gcl/suprv1.o\" -w") returned a
non-zero value 0.

Fast links are on: do (use-fast-links nil) for debugging
Broken at UNLESS.  Type :H for Help.
1 (Continue) Continues anyway.
2 (Abort) Return to top level.

GCC apparently doesn't like this line:

I have looked at the generated C code and I can't tell what part of
suprv1.lisp corresponds to the offending C code.

I can coax GCL into completing the build by responding "yes"
to some questions (from defsystem, I think). With that image
I can execute :lisp (compile-file "src/suprv1.lisp") and it succeeds.
The generated C code is somewhat different -- in particular the
bit about writec_stream doesn't appear.

suprv1.lisp was recently modified to remove stale conditionalizations
and various other minor changes. Those changes shouldn't have
changed any observable behavior.

I wonder if someone familiar with GCL can take a look at
suprv1.lisp and see what might trigger the problem.
I can email the generated C code if anyone is interested.

Thanks for any light you can shed on this problem.

Robert Dodier

reply via email to

[Prev in Thread] Current Thread [Next in Thread]