Hi, On Monday 06 September 2004 15:43, Dave Neary wrote: > Hi all, > > It seems there was a problem with my last mail, so here it is again. > > A colleague of mine, Philippe Couspeyre, and I (but mostly Philippe) > ported sane (and the net back-end, notably) to Win32 natively, using > winsock, to avoid a dependency on cygwin for JNI. This is, as far as I > know, Philippe's first free software contribution :) > > Here's a diff against a post-configured sane 1.0.14, gzipped because I > noticed it was quite large. > > Sorry it's not cleaner, but our priority was to get something working, > so it's not even certain that the same sources will compile on *nix. We > did try to group win32 specific changes together to minimise #ifdefs, > though.
while having a closer look to that patch, I'd say that it is more or less useless for integration - sorry. A patch that will make sane compilation work on a Win32 box natively should bury all the magic within the configure stuff. A post-configured sane diff is not what we can integrate. Ain't it possible to cleanup this patch? Ciao, Gerhard
