Summary: | possible broken code generation (gcc-3.4.1, c++) | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | eleven |
Component: | [OLD] GCC Porting | Assignee: | Please assign to toolchain <gcc-porting> |
Status: | RESOLVED NEEDINFO | ||
Severity: | major | ||
Priority: | High | ||
Version: | unspecified | ||
Hardware: | AMD64 | ||
OS: | Linux | ||
URL: | http://ludojad.itpp.pl/~eleven/timest-geborken.cpp | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- |
Description
eleven
2004-08-06 07:30:18 UTC
start: 2004060112:30:01 stomp: 2004060112:33:01 1086111181 1086111001 i see the expected 180 seconds this is with 3.4.1-r2 can you try 3.4.1-r2 to see if you get the same wrong results, although if you did i couldnt offer an explanation since it works for me (i'm on x86, not amd64) I haven't tried it yet, but - adding -Os to gcc generates proper result... more information: 1. it's C++ independent - happens the same after I rewrote it to lame C (timest-geborken.c in the same place as the original). forcing 32bit code with -m32 generates proper code (with gcc 3.4.1): $ gcc -m32 -Wall -o ctest test.c $ ./ctest 2004060112:30:01 2004060112:33:01 1086085981 1086085801 $ gcc -Wall -o ctest test.c $ ./ctest 2004060112:30:01 2004060112:33:01 1086085981 1086089401 ayanami root # g++ timest-geborken.cpp -o test ayanami root # ./test start: 2004060112:30:01 stomp: 2004060112:33:01 1086107581 1086107401 ayanami root # gcc --version gcc (GCC) 3.4.1 (Gentoo Hardened Linux 3.4.1, ssp-3.4-2, pie-8.7.6.3) it seems to work here? and just in case you're curious: ayanami root # /lib/libc.so.6 GNU C Library 20040808 release version 2.3.4, by Roland McGrath et al. Link is dead, so I can get the testcase. Please try with gcc-3.4.4-r1 and reopen if its still an issue. |