]> git.ipfire.org Git - thirdparty/squid.git/log
thirdparty/squid.git
23 years agoread_ahead_gap config directive by Jeffrey D. Wheelhouse. Allows
hno [Fri, 21 Jun 2002 18:58:20 +0000 (18:58 +0000)] 
read_ahead_gap config directive by Jeffrey D. Wheelhouse. Allows
the read-ahead gap to be configured from squid.conf (previously
hardcoded at 16 KB)

23 years agoSome rewriting inspired by suggestions of overflow bugs and other problems
wessels [Thu, 20 Jun 2002 04:54:14 +0000 (04:54 +0000)] 
Some rewriting inspired by suggestions of overflow bugs and other problems
with this code.

denyusers.c and allowusers.c have a fair amount of nearly identical code.
I moved these functions to a new file, called usersfile.[ch].  I didn't
really want to add another source file, but....

23 years agoSome rewriting inspired by suggestions of overflow bugs and other problems
wessels [Thu, 20 Jun 2002 04:53:56 +0000 (04:53 +0000)] 
Some rewriting inspired by suggestions of overflow bugs and other problems
with this code.

denyusers.c and allowusers.c have a fair amount of nearly identical code.
I moved these functions to a new file, called usersfile.[ch].  I didn't
really want to add another source file, but....

The list of allowed/denied users were formerly stored as whitespace
separated tokens in a single char buffer.  The code used strstr() to
search for matches.  This was awkward because the search key had to
be surrounded by space characters.  The new code uses and array
of char buffers -- one for each name.  The new code uses qsort() and
bsearch().

The old code was more robust with the input file.  You could put multiple
names on a single line.  Due to the way the names were stored, it didn't
matter.  The new code is less robust.  It requires one name per line,
and no leading whitespace.

Names longer than NAMELEN are now truncated.

Added an openlog() call and Removed LOG_USER from syslog() calls.

No longer use scanf() in some places.

There is some weird logic about what to do if an allow/deny file
does not exist, or exists but cannot be read, or was read once but
cannot be read later.  I hope it still works the same.

replaced some strcpy() calls with memset() and strncpy().

Fixed a strange usage of ServerArray[] that started at index 1?

#ifdefd-out some debugging code that included fprintfs to stderr, but
was apparently never called.

Rewrote some signal handling stuff to use sigaction() (and fall back
to signal()) and removed calls to sigvec().

23 years agoSmall fixes and clarifications to the cbdata documentation
hno [Wed, 19 Jun 2002 12:29:45 +0000 (12:29 +0000)] 
Small fixes and clarifications to the cbdata documentation

23 years agoCorrected some peerDigest internal function prototypes, prototyped with
hno [Wed, 19 Jun 2002 04:52:32 +0000 (04:52 +0000)] 
Corrected some peerDigest internal function prototypes, prototyped with
a int argument, defined with a ssize_t argument...

23 years agobuffered_log commend update by Guido
hno [Tue, 18 Jun 2002 16:54:17 +0000 (16:54 +0000)] 
buffered_log commend update by Guido

23 years agoBootstrapped
hno [Tue, 18 Jun 2002 06:10:43 +0000 (06:10 +0000)] 
Bootstrapped

23 years agoAdded support for sys/bitypes.h, apparently needed for some of the bittypes
hno [Tue, 18 Jun 2002 00:28:12 +0000 (00:28 +0000)] 
Added support for sys/bitypes.h, apparently needed for some of the bittypes
on tru64 and possibly others.

23 years agoFixed some include dependencies
hno [Tue, 18 Jun 2002 00:27:30 +0000 (00:27 +0000)] 
Fixed some include dependencies

23 years agoFixed that misgiving comment again...
hno [Mon, 17 Jun 2002 02:10:29 +0000 (02:10 +0000)] 
Fixed that misgiving comment again...

23 years agoAdded some functions for accessing HTTP list header values
hno [Mon, 17 Jun 2002 02:05:53 +0000 (02:05 +0000)] 
Added some functions for accessing HTTP list header values

extern String httpHeaderGetListMember(const HttpHeader * hdr, http_hdr_type id, const char *member, const char separator);

extern String httpHeaderGetByNameListMember(const HttpHeader * hdr, const char *name, const char *member, const char separator);

23 years agoForgot a small piece of the auth cleanup patch from external_acl.
hno [Mon, 17 Jun 2002 01:56:08 +0000 (01:56 +0000)] 
Forgot a small piece of the auth cleanup patch from external_acl.

23 years agoCorrected a misgiving comment. httpHeaderGetByName returns by value,
hno [Mon, 17 Jun 2002 01:48:04 +0000 (01:48 +0000)] 
Corrected a misgiving comment. httpHeaderGetByName returns by value,
not pointer.

23 years agoOops.. got the auth logics slightly wrong..
hno [Mon, 17 Jun 2002 01:32:01 +0000 (01:32 +0000)] 
Oops.. got the auth logics slightly wrong..

23 years agoGeneralized the proxy_auth ACL processing one small step further to make
hno [Sun, 16 Jun 2002 23:46:25 +0000 (23:46 +0000)] 
Generalized the proxy_auth ACL processing one small step further to make
it easier for other ACL types to use/require authentication.

23 years agoPrevent logging of "blank" user names. If the username is "" then log
hno [Sun, 16 Jun 2002 17:24:00 +0000 (17:24 +0000)] 
Prevent logging of "blank" user names. If the username is "" then log
as a dash, assuming the "" usename is a false value from ident or partial
authentication.

23 years agoBugzilla #358: Make whois replies cacheable
hno [Sat, 15 Jun 2002 01:31:22 +0000 (01:31 +0000)] 
Bugzilla #358: Make whois replies cacheable

23 years agoBugzilla #351: ignore-reload should override max-age=0
hno [Sat, 15 Jun 2002 01:26:43 +0000 (01:26 +0000)] 
Bugzilla #351: ignore-reload should override max-age=0

23 years agoAllow for half closed pipelines of requests
hno [Sat, 15 Jun 2002 00:45:16 +0000 (00:45 +0000)] 
Allow for half closed pipelines of requests

23 years agoArgh!
hno [Fri, 14 Jun 2002 23:51:39 +0000 (23:51 +0000)] 
Argh!

23 years agoOops.. syntax error when adding the -T client option
hno [Fri, 14 Jun 2002 23:50:07 +0000 (23:50 +0000)] 
Oops.. syntax error when adding the -T client option

23 years agoCorrected parsing of the max_user_ip -s option (forgot the number of allowed
hno [Fri, 14 Jun 2002 22:27:13 +0000 (22:27 +0000)] 
Corrected parsing of the max_user_ip -s option (forgot the number of allowed
IPs)

23 years agoDon't fill up cache.log about max_user_ip. It does work.
hno [Thu, 13 Jun 2002 23:08:01 +0000 (23:08 +0000)] 
Don't fill up cache.log about max_user_ip. It does work.

23 years agoDocumented the max_user_ip -s option for strict mode
hno [Thu, 13 Jun 2002 23:04:18 +0000 (23:04 +0000)] 
Documented the max_user_ip -s option for strict mode

23 years agoBugzilla #361: client -T opion was not implemented
hno [Thu, 13 Jun 2002 14:41:01 +0000 (14:41 +0000)] 
Bugzilla #361: client -T opion was not implemented

23 years agoBootstrapped
hno [Thu, 13 Jun 2002 06:10:44 +0000 (06:10 +0000)] 
Bootstrapped

23 years agoclear out new auth_user_request objects in case of reused memory
robertc [Wed, 12 Jun 2002 17:47:59 +0000 (17:47 +0000)] 
clear out new auth_user_request objects in case of reused memory

23 years agoUpdate Makefile.am syntax
robertc [Wed, 12 Jun 2002 17:45:20 +0000 (17:45 +0000)] 
Update Makefile.am syntax

23 years agoGet rid of the cbdataInternalLock/Unlock requirement. Still a bit confused
hno [Fri, 7 Jun 2002 00:44:35 +0000 (00:44 +0000)] 
Get rid of the cbdataInternalLock/Unlock requirement. Still a bit confused
in how the peer digests and fetch pulls together however, and there is more
work in cleaning this up to make more proper use of cbdata.

23 years agoRenamed httpSendRequestEntry to httpSendRequestEntity
hno [Thu, 6 Jun 2002 21:11:31 +0000 (21:11 +0000)] 
Renamed httpSendRequestEntry to httpSendRequestEntity

23 years agoFix a minor compiler warning
hno [Thu, 6 Jun 2002 21:11:01 +0000 (21:11 +0000)] 
Fix a minor compiler warning

23 years agoOops.. missed a t in \t
hno [Thu, 6 Jun 2002 20:50:15 +0000 (20:50 +0000)] 
Oops.. missed a t in \t

23 years agoTLS encryption support by Michael Cunningham
hno [Tue, 4 Jun 2002 22:05:52 +0000 (22:05 +0000)] 
TLS encryption support by Michael Cunningham

23 years agoUpper the limit on HTTP ports to 128 (was 12)
hno [Tue, 4 Jun 2002 21:04:05 +0000 (21:04 +0000)] 
Upper the limit on HTTP ports to 128 (was 12)

23 years agoBarf if there is too many https_port defined
hno [Tue, 4 Jun 2002 20:42:32 +0000 (20:42 +0000)] 
Barf if there is too many https_port defined

23 years agoUpdated Douch error pages by Bart Koelman
hno [Fri, 31 May 2002 21:17:17 +0000 (21:17 +0000)] 
Updated Douch error pages by Bart Koelman

23 years agonew config.guess, dated 2002-03-04, from autoconf-2.53
wessels [Thu, 23 May 2002 23:58:04 +0000 (23:58 +0000)] 
new config.guess, dated 2002-03-04, from autoconf-2.53

23 years agoBootstrapped
hno [Thu, 23 May 2002 20:03:43 +0000 (20:03 +0000)] 
Bootstrapped

23 years agoThe winbind helper needs XTRA_LIBS, just as mostly everything else...
hno [Thu, 23 May 2002 20:01:19 +0000 (20:01 +0000)] 
The winbind helper needs XTRA_LIBS, just as mostly everything else...
(-lsocket on Solaris and many other platform dependencies are herein)

23 years agocomplain if kill() (as in --enable-kill-parent-hack) returns an error.
wessels [Thu, 23 May 2002 02:48:14 +0000 (02:48 +0000)] 
complain if kill() (as in --enable-kill-parent-hack) returns an error.

23 years agoBootstrapped
hno [Mon, 20 May 2002 08:05:32 +0000 (08:05 +0000)] 
Bootstrapped

23 years agoMissing files added to "make dist"
hno [Mon, 20 May 2002 08:03:55 +0000 (08:03 +0000)] 
Missing files added to "make dist"

23 years agoBootstrapped
hno [Mon, 20 May 2002 07:48:02 +0000 (07:48 +0000)] 
Bootstrapped

23 years agodist winbind files
hno [Mon, 20 May 2002 07:42:50 +0000 (07:42 +0000)] 
dist winbind files

23 years agodist winbind files
hno [Mon, 20 May 2002 07:42:50 +0000 (07:42 +0000)] 
dist winbind files

23 years agoNeed to dist some header files
hno [Mon, 20 May 2002 07:39:53 +0000 (07:39 +0000)] 
Need to dist some header files

23 years agoBootstrapped
hno [Mon, 20 May 2002 07:18:11 +0000 (07:18 +0000)] 
Bootstrapped

23 years agoSupport out-of-tree "make dist" (autoconf.h)
hno [Mon, 20 May 2002 07:16:41 +0000 (07:16 +0000)] 
Support out-of-tree "make dist" (autoconf.h)

23 years agoBootstrapped
hno [Mon, 20 May 2002 05:53:27 +0000 (05:53 +0000)] 
Bootstrapped

23 years agowinbind helper
hno [Mon, 20 May 2002 05:52:56 +0000 (05:52 +0000)] 
winbind helper

23 years agoBootstrapped
hno [Mon, 20 May 2002 05:48:19 +0000 (05:48 +0000)] 
Bootstrapped

23 years agoOops..
hno [Mon, 20 May 2002 05:46:31 +0000 (05:46 +0000)] 
Oops..

23 years agoSamba Winbind ntlm and basic authentication helpers by Francesco Chemolli
hno [Mon, 20 May 2002 05:42:39 +0000 (05:42 +0000)] 
Samba Winbind ntlm and basic authentication helpers by Francesco Chemolli

23 years agoCOSS fixes by Guido.
hno [Sun, 19 May 2002 22:40:56 +0000 (22:40 +0000)] 
COSS fixes by Guido.

Also triggered a small cleanup in the othere store implementations to
get rid of a unneccesary CYGWIN dependency. (binary file mode)

23 years agoBootstrapped
hno [Sun, 19 May 2002 21:16:37 +0000 (21:16 +0000)] 
Bootstrapped

23 years agoNTLMSSP renamed to SMB.
hno [Sun, 19 May 2002 21:07:55 +0000 (21:07 +0000)] 
NTLMSSP renamed to SMB.

Note: The CVS files have been moved. If going back to a older version than
this then one will have to manually move them back to get a checkout that
reflects the older layout.

23 years agocrypt.h is only needed by some helpers, not the Squid core.
hno [Sun, 19 May 2002 20:32:09 +0000 (20:32 +0000)] 
crypt.h is only needed by some helpers, not the Squid core.

Including it in squid.h only increases compile time and causes
portability issues and conflicts.

23 years agoFix a cbdata reference error in coss async_io implementation.. the pointer
hno [Sun, 19 May 2002 20:14:52 +0000 (20:14 +0000)] 
Fix a cbdata reference error in coss async_io implementation.. the pointer
was cleared before being checked for validity.

23 years agoTwo fixes from Steve Bennett s.bennett at lancaster.ac.uk
wessels [Thu, 16 May 2002 01:08:34 +0000 (01:08 +0000)] 
Two fixes from Steve Bennett s.bennett at lancaster.ac.uk

1) The running averager intAverage() approaches zero due to
   roundoff errors when the input values are less than the
   averaging interval, which is happening for redirector
   service times.  For helper processes, we'll store
   the average service time in microseconds now.

2) the cachemgr output shows 'Time' as large numbers for idle
   helpers.  This patch makes use of the 'answer_time' field
   and uses that for calculating service time in cachemgr stats
   output.

23 years agoDocumented the -o option
hno [Wed, 15 May 2002 18:03:57 +0000 (18:03 +0000)] 
Documented the -o option

23 years agoThe fatal helper queue overflow condition should not be limited by
hno [Wed, 15 May 2002 16:30:27 +0000 (16:30 +0000)] 
The fatal helper queue overflow condition should not be limited by
the warning rate limitation.

23 years agoOne quite obvious use of cbdataReferenceValidDone()
hno [Sat, 11 May 2002 00:56:45 +0000 (00:56 +0000)] 
One quite obvious use of cbdataReferenceValidDone()

23 years agoBootstrapped
hno [Tue, 7 May 2002 06:10:49 +0000 (06:10 +0000)] 
Bootstrapped

23 years agoDon't dist autoconf.h (merged from Squid-2.5, thanks Robert)
hno [Mon, 6 May 2002 12:25:14 +0000 (12:25 +0000)] 
Don't dist autoconf.h (merged from Squid-2.5, thanks Robert)

23 years agominimum_icp_query_timeout needs to set icp_query_min, not max..
hno [Fri, 3 May 2002 20:25:29 +0000 (20:25 +0000)] 
minimum_icp_query_timeout needs to set icp_query_min, not max..
thanks to  Evgeny Kotsuba for spotting the error.

23 years agoConverted htcp to use locally created MemoryPool:s rather than using
hno [Tue, 30 Apr 2002 13:59:49 +0000 (13:59 +0000)] 
Converted htcp to use locally created MemoryPool:s rather than using
the MEM_* approach. This to solve reconfigure and not having htcp configured
in a nice manner.

23 years agoUse FD_READ_METHOD rather than direcly calling read()
hno [Sun, 28 Apr 2002 20:55:40 +0000 (20:55 +0000)] 
Use FD_READ_METHOD rather than direcly calling read()

23 years agoDon't overuse bitfields when storing enums.. care must be taken for
hno [Sun, 28 Apr 2002 18:18:07 +0000 (18:18 +0000)] 
Don't overuse bitfields when storing enums.. care must be taken for
signed/unsigned etc. For fields where memoryusage isn't a significant
impact it is better to not care and let the compiler assign field size.

23 years agoBootstrapped
hno [Sun, 28 Apr 2002 06:10:43 +0000 (06:10 +0000)] 
Bootstrapped

23 years agoAssert that there is at least one port defined when trying to figure out
hno [Sun, 28 Apr 2002 04:54:49 +0000 (04:54 +0000)] 
Assert that there is at least one port defined when trying to figure out
out port number..

23 years agobugfix: Initialize pendingfds to zero
hno [Sat, 27 Apr 2002 14:47:43 +0000 (14:47 +0000)] 
bugfix: Initialize pendingfds to zero

23 years agoCheck for select(), not all OS:es have poll...
hno [Sat, 27 Apr 2002 14:46:44 +0000 (14:46 +0000)] 
Check for select(), not all OS:es have poll...

23 years agoftp_sanitycheck option changelog
hno [Tue, 23 Apr 2002 02:38:36 +0000 (02:38 +0000)] 
ftp_sanitycheck option changelog

23 years agoGopher improvements
hno [Mon, 22 Apr 2002 04:47:59 +0000 (04:47 +0000)] 
Gopher improvements

23 years agoconst cleanup
hno [Mon, 22 Apr 2002 04:14:08 +0000 (04:14 +0000)] 
const cleanup

23 years agoMake delay pools less greedy, to only ask for the amount there is
hno [Mon, 22 Apr 2002 03:54:40 +0000 (03:54 +0000)] 
Make delay pools less greedy, to only ask for the amount there is
clients asking for.

23 years agoDon't mess with store client internals.. not unless debugging is
hno [Mon, 22 Apr 2002 03:54:03 +0000 (03:54 +0000)] 
Don't mess with store client internals.. not unless debugging is
enabled anyway.

23 years agoRemoved unused storeClientWaiting function
hno [Mon, 22 Apr 2002 03:52:47 +0000 (03:52 +0000)] 
Removed unused storeClientWaiting function

23 years agoCleanup of our Gopher client to protect from buffer overflows and
hno [Mon, 22 Apr 2002 03:23:15 +0000 (03:23 +0000)] 
Cleanup of our Gopher client to protect from buffer overflows and
to correct the functionality, plus some new pieces that was missing
like the ability to deal with Gopher Info entries and WWW links

23 years agoRewrite of FTP directory parsing to strengthen against possible buffer
hno [Sun, 21 Apr 2002 22:20:25 +0000 (22:20 +0000)] 
Rewrite of FTP directory parsing to strengthen against possible buffer
overflows

23 years agoftp_sanitycheck option (default on) to make Squid sanity check the FTP
hno [Sun, 21 Apr 2002 20:07:05 +0000 (20:07 +0000)] 
ftp_sanitycheck option (default on) to make Squid sanity check the FTP
data connection.
  * Ignore "BAD" PASV replies, asking Squid to connect to another
    server than requested.
  * Ignore PORT and default connections coming from another address
    than expected.

23 years agoBootstrapped
hno [Sun, 21 Apr 2002 06:10:40 +0000 (06:10 +0000)] 
Bootstrapped

23 years agoAC_CHECK_SYSTYPE, a custom macro extending AC_CHECK_TYPE to look in a more
hno [Sat, 20 Apr 2002 16:50:49 +0000 (16:50 +0000)] 
AC_CHECK_SYSTYPE, a custom macro extending AC_CHECK_TYPE to look in a more
suitable set of system headers (same list as AC_CHECK_SIZEOF_SYSTYPE)

23 years agoBootstrapped
hno [Sat, 20 Apr 2002 06:10:40 +0000 (06:10 +0000)] 
Bootstrapped

23 years agoMove radix internal stuff into radix.c, to avoid polluting the namespaces
hno [Sat, 20 Apr 2002 04:23:01 +0000 (04:23 +0000)] 
Move radix internal stuff into radix.c, to avoid polluting the namespaces
of radix.h users (asn.c)

23 years agoExplicitly truncate the value to indicate that we know truncation is
hno [Sat, 20 Apr 2002 03:03:48 +0000 (03:03 +0000)] 
Explicitly truncate the value to indicate that we know truncation is
needed when assigning an int derived value to a char..

23 years ago<inttypes.h> seems to be the proper header for uint16_t etc not <stdint.h>,
hno [Sat, 20 Apr 2002 02:54:19 +0000 (02:54 +0000)] 
<inttypes.h> seems to be the proper header for uint16_t etc not <stdint.h>,
at least according to SUSv2.

23 years agoCleanup of code not usually compiled
hno [Fri, 19 Apr 2002 14:31:21 +0000 (14:31 +0000)] 
Cleanup of code not usually compiled
By Guido

23 years agoSyntactic restructuring of commSetTimeout to make the function more
hno [Fri, 19 Apr 2002 12:26:26 +0000 (12:26 +0000)] 
Syntactic restructuring of commSetTimeout to make the function more
obvious, oblivating the need of a comment explaining what it going on.

23 years agoOops.. deny_info_url introduces a completely wild array dereference
hno [Thu, 18 Apr 2002 22:18:04 +0000 (22:18 +0000)] 
Oops.. deny_info_url introduces a completely wild array dereference

23 years agoThe cbdata update broke extension of comm timeouts
hno [Thu, 18 Apr 2002 22:09:37 +0000 (22:09 +0000)] 
The cbdata update broke extension of comm timeouts

23 years agoprintf format bugfixes to the printf format bugfixes
hno [Thu, 18 Apr 2002 22:08:21 +0000 (22:08 +0000)] 
printf format bugfixes to the printf format bugfixes

23 years agoOops.. almost got it right.
hno [Thu, 18 Apr 2002 04:09:39 +0000 (04:09 +0000)] 
Oops.. almost got it right.

23 years agoauthenticate_ip_ttl_is_strict comment
hno [Thu, 18 Apr 2002 04:08:02 +0000 (04:08 +0000)] 
authenticate_ip_ttl_is_strict comment

23 years agoCorrected a merge conflict from the cbdata branch.
hno [Thu, 18 Apr 2002 03:53:26 +0000 (03:53 +0000)] 
Corrected a merge conflict from the cbdata branch.

23 years agoThese comments before aclIpNetworkCompare() turned out to be true:
wessels [Wed, 17 Apr 2002 11:31:46 +0000 (11:31 +0000)] 
These comments before aclIpNetworkCompare() turned out to be true:

  * NOTE: this is very similar to aclIpNetworkCompare and it's not yet
  * clear whether this OK. The problem could be with when a network
  * is a subset of the other networks:
  *
  * 128.1.2.0/255.255.255.128 == 128.1.2.0/255.255.255.0 ?
  *
  * Currently only the first address of the first network is used.

The aclIpNetworkCompare() function did not detect collisions and/or
overlapping addresses that can confuse the splay sorting algorithm.
This was proven with an ACL like:

acl a src 1.2.3.4/32
acl a src 1.2.3.0/24

...and then testing the access controls with this sequence of
source IP addresses:

9.9.9.9 correctly denied
1.2.3.4 correctly allowed
1.2.3.5 incorrectly denied
1.2.3.4 correctly allowed
1.2.3.3 correctly allowed
1.2.3.5 correctly allowed

This patch creates two functions for use by the splay library.  One
is used for inserting new ACL entries.  It complains when it detects
a collision/overlap.  The other is used for checking the access
control lists.

I also discovered that we were technically passing the wrong data
type to aclIpNetworkCompare() from aclMatchIp() (via the splay
routines).  The first argument was a 'struct in_addr' but should
really be a 'struct acl_ip_data'.  There was no harm, apparently,
because the first element of acl_ip_data is an in_addr, and the
only member that aclIpNetworkCompare() accesses.  Perhaps this was
intentional, but I doubt it.

23 years agoadd *.gif to .cvsignore
wessels [Wed, 17 Apr 2002 05:13:06 +0000 (05:13 +0000)] 
add *.gif to .cvsignore

23 years agoadd the binaries to .cvsignore
wessels [Wed, 17 Apr 2002 05:12:42 +0000 (05:12 +0000)] 
add the binaries to .cvsignore

23 years agoadding .cvsignore files
wessels [Wed, 17 Apr 2002 05:08:44 +0000 (05:08 +0000)] 
adding .cvsignore files

23 years agooops, undo this mistakenly-committed change
wessels [Wed, 17 Apr 2002 04:45:25 +0000 (04:45 +0000)] 
oops, undo this mistakenly-committed change