Danny, according to that bug report in SF, you said that you're not
submitting that bug to GCC because you couldn't test it on Win 2K and
XP. Now it's been tested on both platforms, are you willing to submit
that patch to GCC?
No, I have an alternative patch that also fixes failure of mingw-hosted
gcc to remove duplicate include paths (because of lack of meaningful
inodes). I will submit that as soon as I can afford to fix some
hardware problems on my gcc development machine. It may be awhile. I
can post the patch to you privately for testing in the meantime.
Danny