---------- Forwarded message ---------- From: <[EMAIL PROTECTED]> Date: Mon, Jun 23, 2008 at 2:48 PM Subject: Re: avoid ssl_eof error on SSL_WANT_READ and friends To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED]
Dear bradfitz, New code review comments by Jacob have been published. Please go to http://codereview.appspot.com/2341 to read them. Message: I've uploaded a new patch Details: http://codereview.appspot.com/2341/diff/1/2 File lib/DJabberd/Stanza/StartTLS.pm (right): http://codereview.appspot.com/2341/diff/1/2#newcode74 Line 74: my $err = Net::SSLeay::get_error($ssl,-1); On 2008/06/19 20:45:39, bradfitz wrote: > space after comma > Done. http://codereview.appspot.com/2341/diff/1/2#newcode85 Line 85: return $err; On 2008/06/19 20:45:39, bradfitz wrote: > Does this function return a boolean or an error code? The name suggests a > boolean, but the return value suggests an error code. > > Both! It returns an error code or zero, and the return value is used as if it were a boolean in the if check in Connection to decide if we should close or not close. If we should close, we want to log the error code, so we log the value returned.... I've changed the function name to actual_error_on_empty_read, is that better? Open to suggestions. Issue Description: Incrementing a counter on empty SSL reads is not the correct way to determine ssl_eof (http://code.sixapart.com/trac/djabberd/changeset/758) Instead we should check the error code to see if it is one of the SSL errors for which we are supposed to retry( http://www.openssl.org/docs/ssl/SSL_get_error.html) Sincerely, Your friendly code review daemon (http://codereview.appspot.com/).