Sorry, it's my fault. I tried again in GCC 3.4.6 in a separate program with -Os, it's a problem. But it's running okay in GCC 4.3.x. So Ian is right, it should be a bug in 3.4.6. Thanks a lot, Win. -----Original Message----- From: gcc-help-owner@xxxxxxxxxxx [mailto:gcc-help-owner@xxxxxxxxxxx] On Behalf Of Tech Win Sent: July 1, 2009 20:09 To: 'Ian Lance Taylor' Cc: gcc-help@xxxxxxxxxxx Subject: RE: What happened when using -Os with gcc? I have to bear with GCC 3.4.6 because I don't have permission to upgrade it. The thing is this piece of code doesn't work if putting it in our system and with the -Os option, if I try it separately, seems okay. Maybe you suspect it's our system's bug, but if I don't use gcc optimization (-O0), or use -O1, -O2, -O3, all no problem. -----Original Message----- From: Ian Lance Taylor [mailto:iant@xxxxxxxxxx] Sent: July 1, 2009 19:34 To: Tech Win Cc: gcc-help@xxxxxxxxxxx Subject: Re: What happened when using -Os with gcc? "Tech Win" <tech_win@xxxxxxx> writes: > But if I use -Os when compiling, the result is: > ================= LAST =============== > 1 --- 1 > 1 --- 0 > 1 --- 0 > t=2525.00 This certainly suggests a bug in the compiler. > My GCC version is 3.4.6. I tried 4.3.2 on i686-unknown-linux-gnu, and it worked as expected with both -O2 and -O2. The current version of gcc is 4.4.0. Ian Checked by AVG - www.avg.com Version: 8.5.375 / Virus Database: 270.13.2/2213 - Release Date: 07/01/09 18:07:00 Checked by AVG - www.avg.com Version: 8.5.375 / Virus Database: 270.13.2/2213 - Release Date: 07/01/09 18:07:00
