From: Oscar F. <of...@wa...> - 2003-09-20 14:50:54
|
"Jeremy Godfrey" <jn...@nt...> writes: > Interesting update: > > I have re-built the test case below using straight MinGW and it is > running fine now. I had been using the version of MinGW that comes > with the latest cygwin and compiling with -mno-cygwin. Jeremy, I was unable to reproduce the problem on Win2K using gcc 3.3.1, MinGW runtime 3.0, MinGW W32API 2.3. My machine is an Athlon 550 MHz with 900 MB RAM. I tried the test case compiled with g++ -g -mthreads and with g++ -g -mthreads -O2 and ran it for 12 minutes each time. It was necessary to add #include <windows.h> after #include <string> Next time you report a problem, please provide at least as much details as shown above. This helps us to help you. I guess the gcc version your cygwin uses has not the MinGW patches, or your MinGW runtime is too old. The fact that your test case doesn't require an #include <windows.h> hints that you are not using the most recent MinGW installation. Earnie & Co.: what's the current policy wrt -mno-cygwin and this mailing list? -- Oscar |