On Mon, Aug 14, 2006 at 06:10:36PM +0100, Dave Korn wrote: > Thanks for the report. This has been mentioned before, it's a problem in >gcc's 'specs', which are command-line pattern matching strings (more-or-less) >used by the gcc compiler driver to select the right options for the >sub-programs that it invokes. It's basically a consequence of the fact that >we're using the same driver to drive a native compiler /and/ a cross-compiler; >normally, a cross-compiler wouldn't include any of the system paths such as >/usr/lib at all, but for a native compiler that's exactly what you want it to >do.
The /usr/lib inclusion is actually likely to be a binutils "problem" since ld includes /usr/lib by default. Possibly adding -Wl,-nostdlib to the command line would help. In any event, I really think that -mno-cygwin is 1) a bad choice for an option name and 2) a bad way to accomplish what is essentially cross-compilation. I regret that I stood by idly and watched that option go into gcc. cgf -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
