dot.{emacs,mosaic-hotlist-default}.
Install sample files into share/examples/w3 instead of to share/doc/w3.
Use PKG_PREFIX in the installation/deinstallation scripts.
Some Makefile cleanup.
Submitted by: (Kiriyama Kazuhiko) <kiri@kiri.toba-cmt.ac.jp>
- Fixed announcement bug. Announcements were always off unless
a file was specified.
- Fixed wrong number of args to examine_select() debug.
- Fixed null-string content-type
- Don't cache replies with 'Set-Cookie:' headers.
- Fixed bug when client issues IMS, Squid has stale object and
Squid's lastmod time is greater than the client IMS time.
A 304 reply would be appropriate for Squid, but not the client
(diagnosed by Mark Treacy).
- Fixed httpBuildRequestHeader() content length bug which breaks for
really large POST requests (Takahiro Yugawa).
- Fixed 'passthrough_proxy' to pick up port number from list of
neighbors.
- Ensure pid file is world-readable if umask is set otherwise
(Doug Urner).
- Collect statistics on a few more HTTP headers.
Reviewed by:
Submitted by:
Obtained from:
This fixes some small bugs, including a CR/LF problem and (I think) two
problems with aborting. The author has hinted that this might be the last
beta before 1.1.0.
Among the changes:
- remove patch-aa, it's now in the standard distribution
- checks for libgnumalloc for <= FreeBSD-2.1.x systems to avoid the
nasty libc malloc problems.
- more useful data in the logs
- support for poll() syscall with unlimited number of file descriptors
if present instead of select()'s restrictions.
(Standard FreeBSD doesn't yet have poll()..)
- the usual bug fixes
PLISTs.
Note: I know that this is going to break some symlinks and/or .so
includes, I will back some of these out as I run into these during
package building.
Included in this update:
- large objects fix, apparently large objects could get stuck (eg:
netscape, msie downloads etc)
- configurable 1st and 2nd level directory hash sizes
- can have a password in squid.conf for cachemgr.cgi
- POST is now passthrough
- other bugfixes
- other new features and corresponding bugfixes :-)
Reviewed by:
Submitted by:
Obtained from:
Don't send RunCache's output to /dev/console when run from the rc.d script,
it's a bit messy and unneeded.
Reviewed by:
Submitted by: asami
Obtained from:
files were repository copied to www/squid11, then this commit brings it
to 1.1.b13. (This was Satoshi's idea :-) It preserves the history)
Squid-1.0 and 1.1 are under parallel development, kinda like when we had
2.1-stable and 2.2-current in parallel development. The 1.0 code is well
polished, and 1.1 is "on the bleeding edge" as such. The features and
performance are much improved, but it can be a bit hair-raising. I
personally have no major hassles with 1.1beta13.
Among the nicer things that this version has over 1.0:
- URL redirector.. ie: you can rewrite url's of sites with "mirrors"
so that you don't have 15 copies of the same files.
- optional ident logging
- improved acl's
- dramatically improved cache directory structure (scales much better with
gigantic disk caches)
- much improved DNS ttl handling (esp. with resolver hack)
- more control over neighbor status; parent, sibling etc.
- much improved refresh rules to help combat stupid sites that needlessly
set the Expires: field to zero when it doesn't need to be. (This is fine
when it's genuinely needed, but some sites really abuse it to to attempt
to negate caching to get inflated hit counts etc)
Reviewed by:
Submitted by:
Obtained from:
and stability measures.
This port installs in a "FreeBSD-native" tree (like apache) rather than
with a mini hierarchy under /usr/local/squid/{bin,etc,cache,logs,...}
(the default behavior seems to have changed between 1.0.0 and 1.0.20)
Also, build a rc.d/squid.sh script.
It is impossible to suppress initial HTML preamble for directories, i.e.
<HEAD><TITLE>Index of dir</TITLE></HEAD><BODY>
It means that it is impossible to change <TITLE> or add any
<META HTTP-EQUIV...> tags to <HEAD> section or change <BODY>
attributes without HTML syntax violation (.asis, cern_meta, etc.
not helps here too).
Fix:
I add "SuppressHTMLPreamble" option to "IndexOptions". When this option
is set _and_ HEADER.html (or what you set as it) is present and readable,
standard <HEAD><TITLE>Index of dir</TITLE></HEAD><BODY> preamble
will be suppressed assuming you have right HTML preamble in your
HEADER.html. It solves all problems mentioned above.
all the COMMENTs! No package names, no version numbers, no "this is
absolutix-3.1.2" type comments that have zero information contents.
Now, without any bad examples to follow, nobody has an excuse to import
a port with those kind of comments. :)
Phew! 238 ports modified!
