Joakim Tjernlund wrote:
You did not confirm if you had the sockets in non-blocking mode or not
(at the time you observed the condition for which the patch improved).
As far as I can tell it is in non blocking mode. Can one have
EAGAIN otherwise?
Well that is a question to ask, if your platform was a lesser known Unix
there are quirks/bugs/features. Other reasons could be related to
restartable system-calls mechanisms with signal handling.
Writing a simple/small test case should not be an issue, the same
application which can be used for server or client end maybe with
different command line options.
Not for me, The SLL stuff is embedded here and there in the webserver
we use (GoAhead) and I don't know enough openSLL to write one.
Then maybe the openssl-users mailing list is your next port of call.
i.e. confirm your usage of OpenSSL is correct.
Sounds like you hacked on OpenSSL to fix some kind of observable
behavior. Your hack might have achieved the desired effect for you but
was probably not the correct fix to the problem.
Darryl
______________________________________________________________________
OpenSSL Project http://www.openssl.org
Development Mailing List openssl-dev@openssl.org
Automated List Manager majord...@openssl.org