Mailing List Archive

errors in 0.9.5
I've built gnupg 0.9.5, and I've noticed the following:
(Linux i386, glibc2, egcs-1.0.2)

With plain ./configure, the compilation goes fine, but make check
fails with
=======================================
PASS: detachm.test
expect: invalid spawn id (7)
while executing
"expect -nobrace -ex "public and secret key created and signed.\r" {
exit 0 } eof { exit 1 } "
invoked from within
"expect {
-ex "public and secret key created and signed.\r" { exit 0 }
eof { exit 1 }
}"
FAIL: genkey1024.test
PASS: conventional.test
========================================

If compiled with
CFLAGS= -m486 -fno-strength-reduce

The compilation stops with this error:
=======================================
gcc -m486 -fno-strength-reduce -Wall -Wcast-align -Wshadow -Wstrict-prototy
pes -o mpicalc mpicalc.o ../cipher/libcipher.a ../util/libutil.a ..
/mpi/libmpi.a ../util/libutil.a -ldl -lnsl -lgdbm -lz
../mpi/libmpi.a(mpi-add.o): In function `mpi_add_ui':
mpi-add.o(.text+0x97): undefined reference to `mpihelp_add_1'
mpi-add.o(.text+0x111): undefined reference to `mpihelp_sub_1'
../mpi/libmpi.a(mpi-add.o): In function `mpi_add':
mpi-add.o(.text+0x32d): undefined reference to `mpihelp_sub'
mpi-add.o(.text+0x515): undefined reference to `mpihelp_add'
...
[cut]
..
mpih-mul.o(.text+0x150a): undefined reference to `mpihelp_add_1'
../mpi/libmpi.a(mpih-mul.o)(.text+0x15c8): more undefined references to `mpihelp
_add_1' follow
collect2: ld returned 1 exit status
=======================================

The CFLAGS "-m486 -fno-strength-reduce" are the standard flags for
Pentium processors set up by rpm.

--
Fabio Coatti http://felix.unife.it/~cova
Ferrara Linux Users Group http://flug.unife.it
GnuPG fp:6AB9 277E 9AA7 9D20 E82C 9EE7 2D17 E351 3DCB 0CDC
Old SysOps never die... they simply forget their password.
Re: errors in 0.9.5 [ In reply to ]
Fabio Coatti <cova@felix.unife.it> writes:

> FAIL: genkey1024.test

$ head README

> If compiled with
> CFLAGS= -m486 -fno-strength-reduce

It seems that the automatic detection of underscores or not make more
problems than the old case statement. Hmmmm...

--
Werner Koch at guug.de www.gnupg.org keyid 621CC013