In message <[EMAIL PROTECTED]> on Fri, 17 Oct 2003 20:20:21 +0200, Stefan Olsson <[EMAIL PROTECTED]> said:Yes, 0.9.7c is the version.
stefan> Hi!
stefan> stefan> I posted this question a couple of weeks ago and have not received any stefan> answer...
stefan> ...maybe because it's a stupid question?
stefan> openssl is compiled using: "./config --prefix=/usr/local stefan> --openssldir=/usr/local/openssl threads shared" and works like a charm, stefan> but the following lines (in eng_openssl.c)
stefan> stefan> #ifdef TEST_ENG_OPENSSL_RC4_P_INIT
stefan> fprintf(stderr, "(TEST_ENG_OPENSSL_RC4) test_init_key() called\n");
stefan> #endif
stefan> stefan> generates unwanted output to the console when a client connects (using stefan> Internet Explorer for instance) to our application. What have we missed? stefan> A compiler flag? Should this be #defined by default?
Which version of OpenSSL are you talking about? In 0.9.7c, the following comment can be found in that file:
Oki. If I don't place that call (or more specific does not specify any engine at all) it works just fine, but our application needs Atalla support for performance and if I don't make a ENGINE_load_openssl(); call before ENGINE_by_id this call failes. What am I missing here?
/* This testing gunk is implemented (and explained) lower down. It also assumes * the application explicitly calls "ENGINE_load_openssl()" because this is no * longer automatic in ENGINE_load_builtin_engines(). */
See the part that says the application explicitely has to call ENGINE_load_openssl() to get that engine loaded at all? are you doing something like that in your application? If so, don't.
Brgds
/Stefan
-- If you make people think they're thinking, they'll love you; but if you really make them think they'll hate you.
smime.p7s
Description: S/MIME Cryptographic Signature
