I just built a version of this test on a linux box and ran it: I get
EINVAL back from every single call to unlock.

Still can't figure out what's causing it.

-----Original Message-----
From: curl-library-bounces_at_cool.haxx.se
[mailto:curl-library-bounces_at_cool.haxx.se] On Behalf Of Michael
Mastroianni
Sent: Tuesday, May 24, 2005 1:10 PM
To: libcurl development
Subject: RE: share interface

Sorry about the curl_easy_init_e--- that's a hack that I did to get ares
failure codes out of it...

I put in a pthreads mutex.... I thought the mutex was necessary to use
this stuff?

If not, go ahead and comment out the guts of the lock methods-- I put in
the printfs.

Thanks for the help.

Regards,
Michael

-----Original Message-----
From: curl-library-bounces_at_cool.haxx.se
[mailto:curl-library-bounces_at_cool.haxx.se] On Behalf Of Daniel Stenberg
Sent: Tuesday, May 24, 2005 11:23 AM
To: libcurl development
Subject: RE: share interface

On Tue, 24 May 2005, Michael Mastroianni wrote:

> Sorry, I was trying to make it easy: how about a header and a cpp
file?

That's a good idea!

... but I don't have windows so it doesn't compile for me. Can't you
just make
the functions write something to stdout instead of doing the critical
section
stuff? Since it is for testing for a libcurl problem we can fake the
locks/unlocks, right?

... and I don't have any curl_easy_init_e() function so it won't link
even
with the issue above fixed.

-- 
  Commercial curl and libcurl Technical Support:
http://haxx.se/curl.html