Figured it out.  It was a bug in Crypt::SSLeay ... not sure how it ever
even worked.  I'm sending them the following patch:

--- SSLeay.xs.orig      2002-08-01 17:43:22.000000000 -0400
+++ SSLeay.xs   2003-01-29 21:41:17.000000000 -0500
@@ -109,6 +109,7 @@
                SSLeay_add_all_algorithms();
                SSL_load_error_strings();
                ERR_load_crypto_strings();
+               SSL_library_init();
               bNotFirstTime = 1;
         }
         RAND_seed(buf,sizeof buf);


Geeze.  Crypt:SSLeay's make test actually passes now, it's amazing.  ;)

--Cliff

Reply via email to