What is the best way to get and maintain a relatively stable version
of 2.5 with NTLM authentication? I don't want to be a developer of the
application, just to use it! I'm setting up a test server and I figure
that 2.5 will be stable by the time I've got it working, so I might as
well start now. More details on what I'm doing and the problems I've
had are below, if anyone is interested.
I am planning to set up a section of a public library for public
access computing; a web browser, NS, MSIE, or both and MS Word, on NT
4.0 workstations with a SAMBA server and Squid on a SPARC 2 processor
RedHat 6.2 server. With the SAMBA server I want NTLM authentication,
along with policies and profiles, to insure that use of the internet
is logged and goes through the SAMBA server. I would like to be able
to extract statistics and restrict usage by time, sites, bandwidth,
file-type access, and bandwidth. I experimented before, using a MySQL
database, authorization written for Apache, and some PERL scripts, but
it was inadequate for what I wanted. (I also experimented with PATC -
which is still hosted on SourceForge but currently dormant - to
provide count-down timers and other facilities that are useful in a
public library.) I would like an integrated single set up where the
ACLs are based on logins; ultimately providing a mix, from anonymous
public logins for 15 minutes to check a hotmail account, to a real
account with a little disk space to someone who wants to work on a
research project.
I realize, that even though I don't think of myself as a developer, I
would have to compile the latest version of 2.5. From posts here, it
seems that RPMs won't do it and that I might need to make a few tiny
changes anyhow.
Sadly, what happened is I got a little confused. I thought I followed
the instructions for anonymous CVS when I tried to get the latest,
more or less stable, development version of 2.5 from
http://squid.sourceforge.net, but I realized something went wrong when
the file 'configure' did not come down with the rest of it. After
looking through the notes of sourceforge I went to
http://cvs.squid-cache.org which looks to be a mirror. I downloaded
squid-head-2000103190000 and I'm compiling as I write this. I worry
that the date being relatively recent, may reflect this as more
experimental than I'm looking for. Also, I don't think getting the
configure file with one method and not another should be the only
guiding factor. I'm not sure if the label 'head' means there is a tail
or body I need from elsewhere. It was a tarball and thus easier for me
to understand what to do.
I'm sure my problem is my lack of understanding of CVS. What should I
do, so I can be using a relatively stable version of 2.5 so I can test
my SAMBA server -Squid Server setup. Is the code on SourceForge a
better choice. Should I be downloading tarballs, through CVS, or
perhaps someone has a small shell script that will keep me current in
relatively staple Squid 2.5s while skipping less stable test versions.
-- Josh Kuperman josh@saratoga.lib.ny.usReceived on Tue Mar 20 2001 - 07:54:09 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:58:44 MST