Open GoogleCodeExporter opened 9 years ago
Hi, I'd be interested in seeing the logs & stack trace.
Maybe I'll also need a network trace, but let's start with the info you already
have.
You can send them privately to lgo@apache.org.
thanks,
Lieven
Original comment by lieven.govaerts@gmail.com
on 19 Aug 2015 at 6:54
Hi Lieven, are you happy with the serf log output and the core file or do you
want to see a Wireshark capture and Apache log file too?
Original comment by 1983-01...@gmx.net
on 19 Aug 2015 at 8:01
Serf log and printed stack trace are fine. A coredump file is not much use to
me because I don't have a FreeBSD system. Thx.
Original comment by lieven.govaerts@gmail.com
on 19 Aug 2015 at 10:42
Appropriate files have been sent.
Original comment by 1983-01...@gmx.net
on 19 Aug 2015 at 1:34
The problem:
The authentication handler initializes the spnego authentication scheme only
for the first connection to a specific server [1], but then uses the authn info
object per connection [2]. That will fail for the 2nd, 3rd ... connection
because there that object wasn't initialised.
[1] auth.c/handle_auth_headers() : authn_info =
serf__get_authn_info_for_server(conn);
[2] auth_spnego.c/serf__handle_spnego_auth() : gss_authn_info_t *gss_info =
(code == 401) ? conn->authn_info.baton
Original comment by lieven.govaerts@gmail.com
on 20 Aug 2015 at 8:02
Lieven, that is obviously a bug isn't it?
This does not cause trouble unless Subversion will start using concurrent
connections, does it?
Original comment by 1983-01...@gmx.net
on 21 Aug 2015 at 10:18
Original issue reported on code.google.com by
1983-01...@gmx.net
on 15 Aug 2015 at 8:56