[Chicken-users] Building Chicken 2.41 fails w/seg-fault on Xenon and PowerPc

[Will Farr]

Hello all,

I've tried to build chicken 2.41 today on a PowerPc G4 and an Intel
Xenon (running, respectively, Mac OS 10.3.9 and Linux
2.4.20---probably from RedHat, but I'm not actually sure) and both
builds have failded with a segfault compiling csi with the newly-built
chicken:

creating chicken
/bin/sh ./libtool --mode=link gcc  -Os -fomit-frame-pointer
-fno-strict-aliasing -Wall -Wno-unused -Wno-uninitialized
-DHAVE_ALLOCA_H -DC_ENABLE_PTABLES  -export-dynamic -o chicken-static
-static   chicken.o support.o easyffi.o compiler.o optimizer.o
c-platform.o c-backend.o batch-driver.o libchicken.la -ldl -lm -ldl
-lm
gcc -Os -fomit-frame-pointer -fno-strict-aliasing -Wall -Wno-unused
-Wno-uninitialized -DHAVE_ALLOCA_H -DC_ENABLE_PTABLES -o
chicken-static chicken.o support.o easyffi.o compiler.o optimizer.o
c-platform.o c-backend.o batch-driver.o -Wl,--export-dynamic
./.libs/libchicken.a -ldl -lm
./chicken csi.scm -quiet -no-trace -optimize-level 2 -include-path .
-output-file csi.c -prologue ./build.scm
gmake[1]: *** [csi.c] Segmentation fault
gmake[1]: Leaving directory `/tmp/chicken-2.41'
make: *** [all] Error 2

Also, the nusery size determination fails with a syntax error on both builds:

checking for best nursery size ...
expr: non-numeric argument
16K ->  ms
expr: non-numeric argument
24K ->  ms
expr: non-numeric argument
32K ->  ms
expr: non-numeric argument
48K ->  ms
expr: non-numeric argument
64K ->  ms
expr: non-numeric argument
128K ->  ms
expr: non-numeric argument
256K ->  ms
expr: non-numeric argument
512K ->  ms
expr: non-numeric argument
1024K ->  ms
the optimal size of the nursery seems to be 1024K

Let me know if I can provide any more useful information about these
failures to aid debugging (perhaps these are related?).

Thanks,
Will