On Sun, Apr 22, 2007 at 12:17:57PM +0200, Thomas Schmitt wrote: > I'll have to correct that via sed. > test -z 1 && for ac_header in dummy > should be a single line repair of > for ac_header in
I don't understand this "repair". test -z 1 will always return "false", because "1" is not a zero-length string. So you might as well just comment out the "for ac_header in dummy" line altogether. As far as the original code goes, I'd have to see the entire "for" loop to know what the actual problem is. Since I don't have access to a Solaris system (with its ancient pre-POSIX /bin/sh), and since dash gives me no errors when I run for ac_header in do echo done that means I don't have any systems that are prehistoric enough to duplicate the bug -- assuming this "for" loop really is the bug at all. It seems more likely that there's really some sort of mistake *before* the "for" loop, which would mean tracking backward through 20000 lines of ./configure code to try to find it... good luck with that. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]