On Wed, Mar 09, 2011 at 08:51:46PM +0100, Robert Wruck wrote: > >Thanks; it is not very critical so don't spend too much effort trying to > >find a way to do so at runtime. We can always use the approach Jeff's > >Makefile patch took to make it safe (and potentially slow) by default on > >all Cygwin while still allowing people on an unaffected version to turn > >the workaround off while building. > > The write() issue will presumably be fixed in the next cygwin > release. I already have a version that works on WinXP where it used > to fail. Cool. Does our Makefile know the cygwin version number via "uname -r"? We could at least tweak the build-time patch to turn it on for the right versions (though we should perhaps wait for the fixed version to be released so we know what it is. :) ). Personally I wouldn't bother much with run-time detection. But maybe cygwin people tend to download binary packages and run them on top of arbitrary versions cygwin? I don't know what's normal. -Peff -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html