This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Bug in current snapshots.
- To: GCC-Bugs <gcc-bugs at gcc dot gnu dot org>
- Subject: Bug in current snapshots.
- From: Karl Günter Wünsch <kgw at cswgmbh dot de>
- Date: Tue, 05 Oct 1999 17:29:15 +0200
- Reply-To: kgw at cswgmbh dot de
Hello again,
sorry for the followup on my own bug report, but I had a sudden idea on
how to track down the described problem. Well the idea worked and now it
looks as if the problem isn't caused by some strange floating point
problem but a code generation problem. I have attached both the
c++-Source and the generated x86-assembler code for the function I have
found to be miscompiled. The problem vanished if I add the variable wh
to the cerr-output statement. In the debugger my fears that wh might be
clobbered were confirmed. If anything else beyond the attached files is
needed to track down this code generation bug, please let me know....
The version that I used for my tests was
Reading specs from
/home/charly/egcs-test/lib/gcc-lib/i686-pc-linux-gnu/2.96/specs
gcc version 2.96 19990920 (experimental)
but the same error is still present in the current snapshot from
19991004.
regards
Karl Günter Wünsch
function.cpp
function.s