Squid-2.5.STABLE4, attempt number 3

Another attempt at getting Squid-2.5.STABLE4 out of the door.

Have been waiting a little to see if anything comes up in the NTLM
field but this has to wait for STABLE5 or 3.0 it seems. It seems
there is NTLM related bugs both in Squid and the helpers, but I have
no possibility to look into more details on this issue at this time.

There has been a few minor fixes in Squid-2.5 in the last weeks
additionally holding off the release, but the sum of fixed bugs is
now way over the threshold to motivate a new STABLE release.

Bugfixes classified as major or meduim severity:

2003-09-01 20:01 (Medium)
Assertion error or segmentation fault if using proxy_auth in
delay_access

2003-09-01 20:01 (Medium)
Segmentation fault if proxy_auth with ntlm used in http_reply_access

2003-08-31 09:31 (Medium)
code 407 instead of 403 for authenticated traffic-shaped user

2003-08-28 22:28 (Medium)
Form POSTing troubles with NTLM authentication or other error
responses

2003-08-06 14:06 (Medium)
assertion failed: client_side.c:1478: "size > 0" when using aufs

2003-08-06 13:06 (Medium)
assertion failed: http.c:869: "-1 == cfd || FD_SOCKET ==
fd_table[cfd].type"

2003-07-16 20:16 (Major)
cbdata.c:186: "c->valid" assertion due to peer digest not found

2003-07-16 13:16 (Major)
Crash after ftpTimeout: timeout in SENT_PASV state

2003-07-11 23:11 (Medium)
Client Socket Buffer leak on reply_body_max_size

2003-07-11 22:11 (Medium)
Forward Host headers in place

2003-07-09 22:09 (Medium)
Memory leak in deny_info TCP_RESET

2003-05-27 07:27 (Medium)
Segmentation fault if more than one custom deny_info message defined

in addition there is a huge amount of minor and cosmetic bugfixes. For
details see http://www.squid-cache.org/Versions/v2/2.5/bugs/

Many thanks to all people who have helped with fixing and reporting
bugs. Also many thanks to the customers of MARA Systems who have been
daring to push Squid-2.5 beyond what it was initially thought of and
that way uncovering some rare corner cases we would not have found
otherwise. My strong feeling is that the upcoming 2.5.STABLE4 will be
the best Squid-2.X release ever.

Most of the now open bugs are touching architectural limitations of
what Squid-2.5 can do, especially in the area of using authentication
and other complex acl controls outside of http_access. Most of these
things was not at all possible in previous versions and should not be
seen as a major drawback of Squid-2.5 but as input on what can be
done better for the next Squid release.

Regards
Henrik
Received on Fri Sep 12 2003 - 15:10:03 MDT