FYI, in my case it made the error more apparent (something like "/usr/bin/tail: no such file or directory"), but no less cryptic. The temporary fix is to rename /bin/HEAD to /bin/HEAD.pl (along with the other two scripts, just for consistency). If it ever gets accepted into libwww, it could be a permanent fix as well.
For reasons outlined earlier, I doubt that would ever happen. But even if it does, we'll still need to deal with systems that have the "old" HEAD script installed for the next decade, probably(*). There are places still using perl5.005, not to mention perl4...
--Chuck
(*) Actually, this could probably work; any *cygwin* system can reasonably be expected to have a recent version of libwww. "legacy" systems that keep the old libwww will typically be unix/linux/etc -- but those platforms don't suffer from a confusion between HEAD and head.
However, it was still bad practice for me to write code that directly calls a platform executable, without using the autoconf machinery to set a variable. I screwed up, so I oughta fix it -- even if perl/libwww removes the conflicting file that highlighted my mistake.
I probably should add a test for 'file' as well -- but that usage in libtool predates me (e.g. everybody else is already calling 'file' directly, so I should be able to do so in win32_libid() as well; if it breaks in win32_libid(), it'll break in hundreds of other places in libtool...) So I'll stick with head, for now...
-- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
