Changelog
+Daniel Stenberg (5 Aug 2008)
+- Yehoshua Hershberg found a problem that would make libcurl re-use a
+ connection with the multi interface even if a previous use of it caused a
+ CURLE_PEER_FAILED_VERIFICATION to get returned. I now make sure that failed
+ SSL connections properly close the connections.
+
Daniel Stenberg (4 Aug 2008)
- Test cases 1051, 1052 and 1055 were added by Daniel Fandrich on July 30 and
proved how PUT and POST with a redirect could lead to a "hang" due to the
o no longer link with gdi32 on Windows cross-compiled targets
o HTTP PUT with -C - sent bad Content-Range: header
o HTTP PUT or POST with redirect could lead to hang
+ o re-use of connections with failed SSL connects in the multi interface
This release includes the following known bugs:
Rob Crittenden, Dengminwen, Christopher Palow, Hans-Jurgen May,
Phil Pellouchoud, Eduard Bloch, John Lightsey, Stephen Collyer, Tor Arntsen,
Rolland Dudemaine, Phil Blundell, Scott Barrett, Andreas Schuldei,
- Peter Lamberg, David Bau, Pramod Sharma
+ Peter Lamberg, David Bau, Pramod Sharma, Yehoshua Hershberg
Thanks! (and sorry if I forgot to mention someone)
* Third version of the patch fixing a failure to chose a proper data
type submitted to the mailing list 2008-08-04.
-
-146 - Yehoshua Hershberg's re-using of connections that failed with
- CURLE_PEER_FAILED_VERIFICATION
-
148 - Introduction of m4/reentrant.m4 is triggering some problems on Solaris
systems. The problem manifests when buildconf runs aclocal, at some point
aclocal fails when using GNU m4 version 1.4.5 it runs out of memory. If
/* perform SSL initialization for this socket */
result = Curl_ssl_connect_nonblocking(conn, FIRSTSOCKET, done);
- if(result)
+ if(result) {
+ conn->bits.close = TRUE; /* a failed connection is marked for closure
+ to prevent (bad) re-use or similar */
return result;
+ }
return CURLE_OK;
}