This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Invalid assembly generation with egcs snapshots and gcc 2.8.0
- To: egcs-bugs at cygnus dot com
- Subject: Invalid assembly generation with egcs snapshots and gcc 2.8.0
- From: Hannu Koivisto <Hannu dot Koivisto at vvf dot fi>
- Date: 18 Feb 1998 00:52:30 +0200
Greetings,
A friend of mine was again testing some perverted template
tricks with gcc 2.8.0/Sparc and noticed that when compiling the
code, assembler was complaining. I decided to test the code with
egcs/Ix86 (with appropriate gas installed) and the result was exactly
the same. Both egcs (at least two most recent snapshots) and gcc
2.8.0 had generated following kind of code:
<futile lines cut away>
main:
<more futile lines cut away>
call make_faceted__H1Zi_RCX01_Q2t10temp_types2ZX01b *1_6t_type
addl $4,%esp
<more futile lines cut away>
.ident "GCC: (GNU) egcs-2.91.08 980214 (gcc-2.8.0 release)"
Parameter to the call would seem to be broken. There may be
other broken lines too, of course, but that seemed to be the first
one. The code, consisting of a header file and a main .cc file,
is in attachments. Also the error output is attached. Just say
g++ -S polyset.cc to get the assembly listing. If the erroneous
listing can't be reproduced, feel free to mail me and I'll
compile it.
Although the problem would seem not to depend on any platform,
the egcs system I tested the code in was Debian 1.3.1 libc5 i586,
egcs 980214 (--enable-shared --with-gnu-as --with-gnu-ld).
//Hannu
polyset.cc
polyset.h
polyset.error