Brandon J Van Every <vanevery@xxxxxxxxxxxxxxxxxxx> writes: > No. What happens in practice, is you have Cygwin vs. MinGW, building > with Cygwin library vs. -fno-cygwin library, and support tools like GNU > Make, bash, etc. all at different version levels. The same is true for > all the underlying open source libraries that a moderately complex > project uses. It's a nightmare. It's all completely broken crap. It > works ok on UNIX because the tools are getting built, tested, and > updated all the time. That testing never happens on Windows, so to > quote Chinua Achebe, "Things Fall Apart." > UNIXen tend to just glare at me when I bring these build issues up. The > reason they glare is they *NEVER* face these build issues themselves. > They just assume I'm some kind of whiner. They assume that UNIX > reproducibility is Cygwin / Mingw reproducibility. It just ain't so. I can certainly agree with that, having dealt with integrating patches for some packages that have been ported to Cygwin. Porting is indeed required; Cygwin is not Unix, and various standard Unix things do not work the way that one would expect. -- Russ Allbery (rra@xxxxxxxxxxxx) <http://www.eyrie.org/~eagle/> _______________________________________________ Autoconf mailing list Autoconf@xxxxxxx http://lists.gnu.org/mailman/listinfo/autoconf
