On Fri, Mar 12, 2004 at 03:33:48PM +0000, Paul Warren wrote:
On Fri, Mar 12, 2004 at 03:28:28PM +0000, Chris Lightfoot wrote:
On Fri, Mar 12, 2004 at 03:20:37PM +0000, Paul Warren wrote:
This email didn't make it onto the mailing list, probably due to the size limit. We're trying to persuade iftop to compile on NetBSD, which like any decent UNIX has its own, uniquely broken pthread implementation. This one doesn't seem to be able to spell "pthread_cancel".
hmm. From the transcript it looks like it's failing to link with the pthread library. The thing to do is probably to figure out how to compile a pthread program (like the test one), perhaps from the pthread(3) man page, then try to figure out why configure isn't figuring it out. Are the NetBSD man pages online?
Hmm. I think previously it was failing to link against loads of things, and now it's just down to that one function, BICBW.
Quite right, I was only reading the top of the log :(
According to the man pages for -current on http://www.netbsd.org/, the implementation supports pthread_cancel and pthread_testcancel, and all you need to do is link with -lpthread. Dunno about older versions though.
Can you try compiling the test program (config/pthread.c) with,
cc -lpthread -g -Wall -o pthread pthread.c
and tell me what you get?