|
Welcome to squidGuard
|
|
|
An ultrafast and free filter, redirector and access controller
for Squid
|
Note1: this page is primarily written for
squidGuard-1.0.x. Though most of it still applies to
squidGuard-0.9.x.
Note2: To print all-in-one go here.
Note3: The FAQ is here.
Note4: The ANNOUNCE
message is here.
The release directory for squidGuard is accessible by both
HTTP
and FTP
The current production release is squidGuard-1.1.1
New: A regularly updated sample
blacklist is now also available.
Se the News and changes and Known bugs and misfeatures sections for more info
on the releases.
squidGuard is a combined filter,
redirector and access controller plugin for Squid. It is
squidGuard can be used to
-
limit the web access for some users to a list of accepted/well
known web servers and/or URLs only.
-
block access to some listed or blacklisted web servers and/or URLs
for some users. **)
-
block access to URLs matching a list of regular expressions or
words for some users. **)
-
enforce the use of domainnames/prohibit the use of IP address in
URLs. **)
-
redirect blocked URLs to an "intelligent" CGI based info
page. **)
-
redirect unregistered user to a registration form.
-
redirect popular downloads like Netscape, MSIE etc. to local
copies.
-
redirect banners to an empty GIF. **)
-
have different access rules based on time of day, day of the week,
date etc.
-
have different rules for different user groups.
-
and much more..
Neither squidGuard nor Squid can be used to
-
filter/censor/edit text inside documents
-
filter/censor/edit embeded scripting languages like JavaScript or
VBscript inside HTML
-
*) 100,000 requests in 10seconds on a 500MHz Pentium with lists
of
-
5900 domains
7880 urls
13780 total
-
100,000 requests in 12seconds on a
500MHz Pentium with lists of
-
5900 domains
200000 urls
205900 total
-
I.e. domain and URL listsizes has
neglectable performance effect..
-
**) squidGuard is not a porn or banner filter/blocker, but it is
very well suited for these
purposes too.
squidGuard has many powerful configuration options that lets you:
-
define different time
spaces based on any reasonable combination
of
-
group
sources (users/clients) into distinct categories like
"managers", "employees", "teachers", "students", "customers",
"guests" etc. based on any reasonable combination of
and optionally link the group to a
given time space
-
positively (within business-hours)
-
negatively (outside leisure-time)
-
group
destinations (URLs/servers) into distinct categories like
"local", "customers", "vendors", "banners", "banned" etc. based on
an unlimited number of unlimited lists of
and optionally link the group to a given time space:
-
positively (within business-hours)
-
negatively (outside leisure-time)
-
rewrite/redirect
URLs based on any reasonable combination of
-
URL replacement with
-
silent squid redirect to a common URL
(redirect "new_url")
-
visible client redirect to a common URL
(redirect "302:new_url") ***)
activated by
and optionally with
-
define access control
lists (acl) based on any reasonable combination
of the definitions above by
-
giving each source (user/client) group
- and optionally:
-
link the acl to a given time space
-
positively (within business-hours)
-
negatively (outside leisure-time)
-
defining a fallback/default
ruleset
-
have selective logging by optional log statements in
the: ****)
and optionally anonymized to protect the individuals
(log anonymous "file")
*) Client access control based on domain name requires enabling
reverse lookups (log_fqdn on) in squid.conf.
**) Client access control based on user id requires enabling
RFC931/ident in squid.conf. Note: The RFC931/ident configuration
is changed in squid-2.2 and the RFC931/ident support is
broken in squid-2.2 at least up to STABLE2. We
currently recommend using squid-2.1.PATCH2
in production if RFC931 is used.
***) Note: Visible redirects (302:new-url) are not
supported by some interim versions of Squid (presumably 1.2-2.0).
****) Note: squidGuard is smart enough to open only one
filedescriptor per logfile (i.e. not necessarily one per log
statement); per spawned process of course. Though logging to too
many different files may exeed your system's concurrent
filedescriptor limit.
squidGuard should compile right out of the box on any modern brand
of UNIX with a development
environment and a recent version (2.X) of the Berkeley DB
library. squidGuard is developed on Sun Solaris-2.6 with gcc-2.7.2.3, bison-1.25, flex-2.5.4. We also test regularly on Linux/RedHat-5.2 with gcc-2.8.1
and Digital
UNIX/OSF1-4.0 with gcc-2.7.2.3. We always use our most recent copy of
the Berkeley DB.
Users have reported success on at least, but not limited to:
-
AIX: 4.1.3, 4.3.2.0/egcs-2.91.66
-
Dec-Unix: OSF1-4.0/gcc-2.7.2.3, 3.2C/gcc-2.7.2.3
-
Linux: RedHat-5.2/gcc-2.8.1
-
Solaris: 2.6/gcc-2.7.2.3
Please report
missing list entries.
The installation guide is in a separate document.
The Configuration is in a separate document.
squidGuard uses Squid's standard
redirector interface. When Squid is used with squidGuard:
-
Squid spawns a configurable number of squidGuard subprocesses at
startup.
-
For each request Squid receives, it sends a line of the format
-
URL ip-address/fqdn ident method\n
through a pipe to a squidGuard subprocesses on its
stdin. ("\n" is the newline character.)
Note: This is after a possible insertion of the
value for "append_domain" by Squid (See
squid.conf).
-
squidGuard checks the received info with the configured access
control lists and decides whether the URL should pass unchanged or
not.
-
squidGuard then sends a line through its stdout via a
pipe back to Squid. The format is
-
\n
or
-
new_URL ip-address/fqdn ident method\n
for passed URLs and modified/blocked URLs respectively. By default
the redirect is done by Squid silently and invisible for the
user/browser. The new URL can be prepended with "302:" to
make the redirect visible for the user by asking the browser to
request the new URL.
Note: A Squid redirector cannot
really block an URL; it must redirect it to something
(preferably an empty image, a stopsign image, a dummy access
denied page or a more or less intelligent and
informative page (CGI)).
-
On a HUP signal Squid terminates all its redirector
subprocesses (squidGuard) and starts new fresh ones.
-
On exit (a TERM signal) Squid terminates all its
redirector subprocesses (i.e. squidGuard) too.
-
A redirector subprocesses (i.e. squidGuard) must never exit while
Squid is running. Though Squid will recover from crashed
redirector subprocess problems with a HUP signal. Anyway
squidGuard has proven rock solid so far.
For each request squidGuard will:
-
try to find a matching client group based on the client IP-address
and optional domainname and user ID information. Note: The
client groups are matched in the order they are defined. Thus a
client group that is a subset of a more general group must come
first of the two to take effect. If the client does not
match a group then the default acl will be used.
Note: The client information must match at least one of
each defined type within the actual group to qualify
(i.e. ip AND domain AND
user).
-
select the corresponding active acl. If no corresponding
acl is active or defined the default acl is selected.
-
try to match the URL to each destination group in the listed order
in the pass rule in the actual acl and for each
destination group in the priority order domainlist,
urllist and expressionlist.
Note: It
is sufficient that the URL matches one of the defined types within
the actual group to qualify (i.e. domainlist OR
urllist OR expressionlist).
-
if a negative group ("!group") is matched,
return the redirect URL for that destination group if
defined or alternatively the redirect URL in the actual
acl if defined or else the redirect URL in the default
acl as the last resort.
-
when a positive group ("group") is matched the stop searching.
-
apply the rewrite rules for the matched destination group if any
and then apply rewrite rules for the acl if any or else
the rewrite rules for the default acl if any.
-
if the URL was changed by a rewrite rule return the
new URL and the suplied information.
-
Otherwise return an empty line indicating no
change to Squid.
The main design goals for squidGuard were, and still are:
-
robust and modular code already from the start
-
good coding standard
-
efficient code and dessign
-
use of existing robust libraries, OS native or free, where
appropriate
The efficiency is achieved by:
-
storing all lists in B-trees.
-
compiling/optimizing all regular expressions at startup.
-
source code in C.
-
using custom-made string comparators where regular expression
match would be an overkill.
-
using a cursor to check the previous key in B-trees to speed up
substring matching.
-
chosing the more efficient system libraries/calls where there are
alternatives.
-
breaking out from the rule checking loops as son as a decision can
be made.
-
avoiding constant evaluation of the rule's time space validity by
using valid/disabled flags. At startup the initial state of the
time dependent rules are evaluated and an alarm is set for the
first moment of change. When the alarm goes off squidGuard
reevaluates all its time dependent rules and sets a new alarm,
etc. etc.
The robustness is achieved by:
-
actively trying to do things the right way already from the start.
-
focusing on and implementing one module at the time.
-
using the well proven Berkeley DB library to achieve much with
simple code.
-
using lex and yacc to make a simple and flexible
configuration file parser that is easy to modify and extend.
-
using gcc -Wall -W... to avoid common errors.
-
publicly reviewable open source code.
Please note:
-
If something is terribly wrong (like syntax error in
squidGuard.conf) squidGuard will fallback to a pass all for
all mode. This is an intensional feature not a bug!
squidGuard can't resume normal operation from this mode without
being restarted by a HUP signal to Squid (not
squidGuard).
-
The client groups are matched in the order they are defined. Thus
a client group that is a subset of a more general group must come
first of the two to take effect.
-
The URL and domain matching is case insensitive. Thus
-
http://www.what.ever.com/where/ever
and
-
http://wWw.wHat.evER.com/Where/Ever
doesn't matter.
-
The protocol (proto://) prefixes are ignored by
squidGuard in both domain and URL matching, but not in the regex
matching.
-
Domain matching includes subdomains. Thus if you have a domainlist
containing bad.com all these URLs will match that
destination group since they are equal to or subdomains of
bad.com:
-
http://bad.com
http://bad.com/whatever
ftp://bad.com
wais://bad.com
http://www2.bad.com
http://whatever.bad.com
http://www56.whatever.bad.com
but not:
-
http://www.verybad.com
unless www.verybad.com or verybad.com is in that
list too.
-
The domain prefixes www[0-9]? and web[0-9]? are
ignored by squidGuard in URL matching. Thus if you have an urllist
containing "bar.com/~baduser" then all these URLs
will match that destination group:
-
http://bar.com/~baduser
http://www.bar.com/~baduser/whatever
http://www2.bar.com/~baduser/whatever/bar.html
but not:
-
http://foo.bar.com/~baduser
unless foo.bar.com/~baduser is in that list too.
-
squidGuard will match subpaths for URLs. Thus if you have an
urllist containing "foo.bar.com/~baduser" then all these
URLs will match that destination group:
-
http://foo.bar.com/~baduser
http://foo.bar.com/~baduser/
http://foo.bar.com/~baduser/index.html
http://foo.bar.com/~baduser/whatever/foo.html
http://www.foo.bar.com/~baduser/whatever
http://www2.foo.bar.com/~baduser/whatever/bar.html
but not:
-
http://foo.bar.com/~baduserhome
nor
-
http://foo.bar.com/~gooduser
http://foo.bar.com/~gooduser/foo.html
http://foo.bar.com/~gooduser/whatever/bar.html
unless foo.bar.com/~baduserhome and/or
foo.bar.com/~gooduser is in that list too.
-
Currently a squidGuard process will reincarnate on a
HUP signal directly to the squidGuard process -- if at
all possible (i.e. execvp() and replace itself with a
fresh squidGuard process with the same command line
arguments). Thus if you suspect a squidGuard process has gone mad,
you don't necessarily need to reconfigure Squid; a
"kill -HUP that_squidGuard_pid" will
probably do.
-
Major news and changes in
-
1.1.1:
-
1.1.0:
-
Brought the doc up to date.
-
1.1.0.beta1:
-
Added support for prebuilt database files. squidGuard will now
fire instantly even with a million entry blacklist, use less RAM
and perform only marginally slower than with the same databases
in-memory-only. This is done transparently in such a way that if
you have the config:
-
dest foo {
-
domainlist foo/domains
}
and "dbhome/foo/domains.db" exists then
"dbhome/foo/domains.db" will be used regardless
if "dbhome/foo/domains" exists or not.
-
"squidGuard -C file" (re)builds file.db
from list file and exits. Note file is
relative to dbhome unless absolute. Also note
"squidGuard -C all" (re)builds databases for
all defined lists.
Note: Prebuilt database are currently only supported for
domainlists and urllists in the destination groups. We plan to
add support for this to the userlist and iplist in the source
groups soon.
-
Added support for runtime updates in prebuilt database files
without signal or restart.
-
"squidGuard -u" updates all defined databases
(file.db) from ditto file.diff where found
and exits. Note file.diff has the simple format:
-
+entry_to_be_added
-entry_to_be_removed
...
(i.e. not the common patch format)
-
Added %p for the path part (without the leading
"/") in the runtime substitution. Thus you may find this urllist
syntax useful for hostredirection:
-
foo.bar http:hotell.server/foo.bar/%p
-
Added support for old K&R compilers through CPP magic inspired
by the original MIT X11 coding standard. So an ANSI-C compiler
is no longer requiered, though gcc is still recommended.
-
Added an up to date 25k entries blacklist; now regularly
updated and published separately in the distribution directory.
-
Contributed the squidGuardRobot. Currently completely
undocumented and not recommended for the average site.
-
Fixes: All known bugs were fixed, including:
-
building with the DB library installed in odd places.
-
building with native lex.
-
userlist exact match bug.
-
should match with the full hostname against domainlist
(i.e. not strip www-prefix etc. here). Thus a domainlist may
now contain a hostname like www.foo.bar if you only
want match with that host and not www2.foo.bar or
whatever.
-
some potential buffer overflows.
-
1.0.0:
-
userlist: Now accepts a file like
/etc/passwd, .htpasswd etc.
-
1.0.0.beta3:
-
1.0.0.beta2:
-
All objects and default ACL: Optional logging of altered
requests (log file)
-
Source objects: User lists in files
(userlist file)
-
Source objects: IP lists in files
(iplist file)
-
CGI: Added a brand new squidGuard.cgi in
samples/ with configurable strings, optional reverse
lookup with autoredirect for in-addr and more.
-
Fixes: All known bugs were fixed.
-
General: Rewritten common logging functions.
-
General: Completed the new comprehensive online
documentation (http://info.ost.eltele.no/freeware/squidGuard/)
-
1.0.0.beta1:
-
New object: time
(time name {...})
-
New built in: in-addr
(pass ... !in-addr ...)
-
All objects and ACLs: Optional time constraints
(... within|outsidename {...} [else {...}])
-
Optimization: Major code review. Now faster than ever in
spite of more functionality.
-
Build: Adapted GNU autoconfigure.
-
Fixes: All known bugs were fixed.
-
General: Established an official home page with new
comprehensive documentation (http://info.ost.eltele.no/freeware/squidGuard/)
Se the change
log for the details and minor changes
We have had squidGuard runing on our Squid proxy for more than a
year now with about 5-11000 listed domains, 6-14000 listed URLs, 3
complex regular expressions and 20-25 acls whereof some with
timeconstrains. So far we haven't seen a single crash, coredump or
hang caused by squidGuard in production. Although squidGuard has
proven rock solid there still are some known minor problems:
-
Not anymore; All known problems have been fixed in
squidGuard-1.1.0.beta1. Though it includes some new
functionality that may have introduced new bugs we haven't
discovered yet.
That is the question.. Here are some possible development issues we
have discussed:
-
Put host addresses in IP lists into B-tree
-
This should dramaticly reduce the performance penalty of
long lists of random client addresses. We should
still keep IP ranges in sequentially linked lists since ranges of
random width are hard to index efficiently.
-
Add syslog support
-
Maybe we should add syslog support as an alternative for the
squidGuard.log.
-
Include support for classless IP ranges in the domainlists
-
We have to find a good storage method and modify the domain
comparator slightly to select IP range mode if the URL host is an
IP address. IP range comparision is allready in use in the client
group match so it shouldn't be to hard.
-
Benchmarking alternative regexp libraries
-
Configure should automaticly select the best regexp library;
native or GNU/FSF.
-
Reconfigure on a HUP signal directly to squidGuard
-
In the beginning where we used only one redirector process, this
seemed like a good idea. I really don't know why one should use
more squidGuard redirectors than CPUs on the system but that seems
to be recommended. So therefor we have got second thoughts about
this since it raises a real challenge to find a way to reconfigure
all the redirectors with one signal. Either this should be done by
having the good Squid guys implement a signal handler for a
dedicated signal that is passed to the redirector processes or
that makes Squid send a HUP signal to its
redirectors. Alternatively this has to be done by some kind of
IPC. Maybe SysV IPC semaphores is an option here, but this is
likely to reduce the portability and increase the complexity of
squidGuard.
-
Preventing users from accessing blocked sites by using their IP
address after a manual nslookup
-
This is partially done by the new optional !in-addr
flag. An alternative path with non blocking reverse lookups before
group matching have been left since it complicates things too much
at the present. An alternative is to use a simple Perl script to
include both the domainname and the IP addresses in the
lists
-
A shared database engine
-
When running multiple squidGuard processes it is of course a waist
of RAM to have each processes hold a complete private copy of the
destination databases in memory. This is now solved with the new
support for prebuilt on-disk databases. Though there are other
possible solutions to avoid this too:
-
Spawn only one squidGuard process. This may be the best solution
on a single CPU system anyway.. And maybe increase squidGuard's
priority (renice(1)) manually if Squid warns about
growing redirector queue.
-
Have a common database engine process with concurrent access by
the squidGuard processes through IPC. This raises a real
challenge to handle problems like:
-
Who should spawn the database engine process and how should it
be done to avoid race conditions?
-
How should the processes comunicate -- through pipes or SysV
IPC?
-
How to implement robustness in case one process in the
"cluster" crashes?
-
What about portability between BSD and SysV systems?
-- Comments and contributions are welcome.
The FAQ is in a separate document.
Here is a guide to related documentation
-
A
happy user's squidGuard page :-)
-
Fabrice
Prigent, Network Administrator at the University of Toulouse,
France, has written an excellent page about squidGuard in both French
and English. He
has done a neutral
comparison of known free redirectors for Squid. He also has
links to some URL/domain
lists/patterns you may find useful in conjunction with Squid
and squidGuard.
-
The
official Squid 1.1 Release Notes:
-
The Squid programmers/authors original documentation on Squid's
redirector interface.
-
The
official Squid redirector FAQ:
-
Frequently Answered Questions about Squid redirectors.
-
The
official Squid ralated software FAQ:
-
Pointers to Squid redirector packages and more.
-
The
official Squid FAQ:
-
Answers to almost all your questions about Squid.
-
Squid
related documentation:
-
Pointers to even more Squid related documentation.
Here is a guide to more or less free software tools you may or may
not need.
-
squidGuard:
-
Well, that is what this is all about :-)
-
Squid:
-
If you don't have Squid you don't need squidGuard either. You may
prefer to download our copy
if you are in Europe or have trouble reaching the Squid home page.
-
libdb.a -
Berkeley DB - The Berkeley Database library version 2.X:
-
squidGuard requires this library. You must install a version 2.X
of DB if your system does not already have it or it has a version
1.X installed (Version 1.X and 2.X may coexist on a system). The
Berkeley Database library can be used free of
charge with squidGuard since squidGuard is GPL. You may
prefer to download our
copy if you are in Europe or have trouble reaching Sleepycat.
Note: All
versions of the Berkeley Database library are called betas. In
contrast to the normal use of the notion beta this does not
mean that the Berkeley Database library is in unstable prerelease
state. Quite contrary, it has proven to be very stable and can
safely be used in production quality software.
-
Gzip -
GNU's replacement for compress:
-
You need gzip/gunzip/zcat
to unzip the squidGuard distribution. If you ever plan to work
with freeware for UNIX-like environments then gzip/gunzip/zcat
is a must. You may prefer to download our copy if you are
in Europe or have trouble reaching The Free Software Foundation or
finding a good mirror.
-
Gmake - GNU make:
-
If you don't have a development environment with a "make"
compatible build tool and stuff listed below, then the Free Software Foundation is your
best friend. The Free Software Foundation makes free versions of
everything you need to set up a development environment for
UNIX-like systems. Besides the University of California,
Berkeley/Berkeley Software
Design, the free software from the Free Software Foundation is the
main reason there are things called 386bsd, NetBSD, FreeBSD,
Linux, etc. etc. Thanks to the constant open source revision and
refinement via the Internet over the years since 1983, the Free
Software Foundation in most cases makes faster, better and less
buggy reincarnations of most UNIX-like tools an libraries than the
commercial vendors do and the code runs on almost any UNIX-like
platform you can think of so you can set up a more uniform
environment on a multivendor fleet. Therefor we recommend using GNU tools wherever you can! You may
prefer to download our
copy if you are in Europe or have trouble reaching The Free Software Foundation or
finding a good mirror.
-
Gcc -
The GNU C Compiler:
-
If you don't have an ANSI C compiler then the Free Software Foundation is your
best friend. Gcc makes faster and less buggy code than many
commercial compilers and runs on almost any UNIX-like platforms
you can think of, so we recommend using it anyway unless your
commercial compiler is really clever. squidGuard is developed
mainly with gcc-2.7.2.3. (The new gcc-2.8.x may be fine too.) You
may prefer to download our copy if you are
in Europe or have trouble reaching The Free Software Foundation or
finding a good mirror.
-
Bison -
The GNU/FSF parser generator, Bison:
-
Byacc
- The Berkeley Yacc parser generator:
-
If you for some strange reason don't have a "yacc" (Yet
Another Compiler Compiler) compatible parser generator then The GNU/FSF parser
generator, (bison) and The Berkeley Yacc
parser generator (byacc) are your friends. You need yacc,
byacc or bison to compile src/sg.y into src/y.tab.c and
src/y.tab.h. Bison makes faster parsers than most commercial
versions of yacc, so we recommend using it anyway as long as its
copyright isn't a problem for you (You may not use Bison to build
commercial software). If you can't use Bison and lack yacc, use
byacc which has a more liberal copyright. (Though the parser speed
doesn't matter much for squidGuard since it is used for parsing
the config file only). You may prefer to download our copies of bison and/or byacc if you
are in Europe or have trouble downloading from the sources.
-
Flex -
The Fast Lexical Analyzer Generator:
-
If you for some strange reason don't have a "lex"
compatible lexical analyzer generator then the Fast Lexical Analyzer Generator
(flex) is your friend. You need lex/flex to compile src/sg.l
into src/lex.yy.c. Flex makes faster scanners than many
(most/all?) commercial versions of lex, so we recommend using it
anyway. (Though the lexer speed doesn't matter much for squidGuard
since it is used for parsing the config file only). squidGuard is
developed mainly with flex-2.5.4a. You may prefer to download our copy if you are
in Europe or have trouble reaching The Free Software Foundation or
finding a good mirror.
-
Regex -
the GNU regex library:
-
Rx - A fast
replacement for the GNU regex library:
-
squidGuard must be linked with a regular expression library that
supports regcomp() and regexec(). If your system
for some strange reason don't have this the GNU regex or rx libraries may be good
alternatives. We haven't done any comparison of these libraries
yet, but have successfully linked squidGuard with both. These may
be faster than many commercial variants anyway, so you may
consider using one of them even if your system have a compatible
library. Finding a fast regular expression executer,
regexec(), may be critical for good performance if you
run squidGuard with many and/or complex regular expressions. The
performance of the regular expression compiler regcomp()
is less critical since all expression compilation is done at load
time, not when handling requests. You may prefer to download our copies if you are
in Europe or have trouble reaching The Free Software Foundation or
finding a good mirror.
-
Apache - The most popular web server
on the Internet:
-
Although your redirectors could point to nonexisting URLs, it
would be a better idea in most configurations to redirect to
existing URLs. So you probably need a web server. If you don't
have a web server we strongly recommend Apache. You can
install it on the proxy server or a nearby server. You may prefer
to download our
copy if you are in Europe or have trouble reaching the apache.org or finding a good mirror.
-
Perl -
The popular all in one language:
-
You don't really need Perl to run squidGuard, but it comes very
handy when making CGI scripts, HTML parsers, URL collectors,
database compactors etc. etc. You may prefer to download our copy if you are
in Europe or have trouble reaching the perl.com or finding a good mirror.
-
Identd
- our free ident daemon
for M$ Windoze:
-
Pidentd - a
free Portable Ident Daemon for Unix:
-
To be able to use user IDs in the
client rules definitions the actual
clients must support RFC931/RFC1413. Pidentd is
for Unix and Identd for
M$ Windoze. Feel free to download our copy of pidentd
or our official distribution of identd.
-
squidlog2combined:
-
squidlog2combined is a free tool that lets you convert native
Squid accesslogs to the Apache combined log format, suitable for
log analyzers like Webalizer and others. It
has some really
nice options.
-
Webalizer -a free web
server log file analyzer:
-
Webalizer is a
fast, free web server log file analysis program. It produces
highly detailed, easily configurable usage reports in HTML format,
for viewing with a standard web browser. With Squid you may find
it usefull in combination with squidlog2combined
and Apache's
logresolve. Feel free to download our copy.
Note: ftp.ost.eltele.no is not
mirroring other sites. We just put a copy there, to have it
handy, of what we manually download and find useful. The copies
are unchanged from the original but may have been compressed with
gzip (.gz) if the original
was compressed with the old compress (.Z).
squidGuard is covered by the GNU General Public
Licence (GPL). The copyright boils down to:
-
You may use squidGuard in both free and commercial services
without fee or registration to ElTele Øst AS.
-
You may freely redistribute unchanged free copies of squidGuard;
precompiled or in source form.
-
There is no warranty for squidGuard.
-
You may not change the copyright.
-
You may not sell or use squidGuard in a commercial software
package without a written agreement with ElTele
Øst AS; and most likely with Sleepycat
too for the DB part.
For more copyright details you have to read the GNU General Public
Licence (GPL).
If you find a bug, have a bug fix, a contribution, a question or a
comment on squidGuard feel free to send it to squidguard@ost.eltele.no
-
The preferred bug fix format is context diff:
-
diff -c origfile fixedfile
-
or
-
diff -cr origdir fixeddir
-
Please explain what problem your diff fixes.
Success reports are very
welcome too.. They makes us even more
enthusiastic!
squidGuard was designed by Pål Baltzersen and
implemented by Lars Erik
Håland
at ElTele Øst
AS and made freely available by permission of ElTele Øst AS.
This page is written and maintained by Pål Baltzersen.
Please report any
errors, broken links, missing stuff, typos, misspellings, wrong
grammar, norwenglish phrases etc. you discover.
-
Thanks to:
-
Gary Lindstrom,
Fabrice Prigent.
Jiri A. Randus.
Joseph Lesko.