Nelson, Thanks for the great reply! This gives me a much better idea of the behavior of the code :)
I erroneously assumed the connection between SECWouldBlock and PR_WOULD_BLOCK_ERROR (I read the notes.txt file a while ago...). Also, a function in the NSS code would in some cases check if a returned result is SECWouldBlock and, if so, would set PR_WOULD_BLOCK_ERROR code. I guess SECWouldBlock would imply PR_WOULD_BLOCK_ERROR, but the converse is not true. It also helped to know the SSL layer offers the same semantics as the underlying NSPR socket - if the NSPR socket is non-blocking, SSL would also offer non-blocking semantics. I.e. if the server_hello message doesn't arrive when the handshake attempts to gather it, the SSL handshake doesn't block on waiting for it, but returns failure and PR_WOULD_BLOCK_ERROR as the error code. Thanks again :) Regards, Peter _______________________________________________ dev-tech-crypto mailing list dev-tech-crypto@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-tech-crypto