"Also sprach David Luyens:"
> Could you share your config with us?

I'm sure I could - but please note that I would be very grateful if you
did NOT quote a 200 line mail of mine (or someone elses) without
commenting on a word of it!  What was the point of attaching the stuff
below (apart from browning me off, showing your subtle disdain for what
had been written, etc?  :( )?

The above 6 words of yours are sufficient and complete in themselves,
surely!  Don't you trust them?


> -----Oorspronkelijk bericht-----
> "Also sprach Alan DeKok:"
> > "Peter T. Breuer" <[EMAIL PROTECTED]> wrote:
> > > get it to work and respond! (I tried gnu-radius and gave up in horror).
> > 
> >   Any particular reason why?
> 
> The configuration was relatively painful! I actually liked freeradius's
> config.  I had great trouble with gnu-radius. It may be the emacs/vi
> users divide.

I recall now that gnu-radius didn't tell me HOW to configure it. While
freeradius nicely put a radiusd.conf file in a etc/raddb directory,
gnu-radius didn't - I had to trace it to see what it was looking for
(etc/raddb/config, if I recall), and then there was nothing anywhere in
the source called *conf* to try and construct a working example from!

The only thing I could see that looked likely was
examples/config.syntax.  And that says up top:

  ## -*- radconf -*-
  ## This file is intended to be an explanation on configuration syntax
  ## issues. It cannot possibly be used as a configuration file.
                ^^^^^^^^^^^^^^^^

Nevertheless, I used it quite happily. The only thing was that it
complained about this line in the example:

 # DBM handling. Can be one of
 #       no      -       Never to use DBM database
 #       yes     -       Use only DBM database and ignore users
 usedbm no;
 ^^^^^^^^^^

saying that the syntax was illegal. Shrug. Also it seemed to be
connected with something called guile, which I didn't really want to
get closer than a bargepole to. There was a whole section devoted to it.

Here's my search for config info:

  src/radius-1.3% find . -name \*conf\*
  ./doc/man/config.so.in
  ./doc/man/config.so
  ./doc/texinfo/config.texi
  ./doc/texinfo/configure.texi
  ./intl/config.charset
  ./raddb/client.conf.in
  ./raddb/config.in
  ./raddb/client.conf
  ./raddb/config
  ./tests/raddb/client.conf.in
  ./tests/raddb/config.in
  ./tests/proxy/client.conf.in
  ./tests/proxy/config.in
  ./tests/atconfig
  ./configure.ac
  ./config.h.in
  ./configure
  ./scripts/config.guess
  ./scripts/config.rpath
  ./scripts/config.sub
  ./scripts/config-conv
  ./scripts/config-conv.awk
  ./radiusd/config.c
  ./radiusd/config.y
  ./radiusd/.deps/config.Po
  ./radiusd/config.o
  ./elisp/radconf-mode.el
  ./db/config.m4
  ./examples/config.syntax
  ./examples/pam.conf
  ./contrib/radsend/paquetes.conf
  ./configure.lineno
  ./config.log
  ./config.status
  ./config.h


Then gnu-radius seemed to somehow pick up the dictionary definitions
from freeradius, possibly through some shared config file. It looked
like gnu-radius perhaps uses all the config file names that are
"deprecated" (but still exist) in freeradius. Anyway, it di8dn't like
any of freeradius' dictionary file syntax, having gone to the trouble
of finding and understanding the pointers to them! What *is* a
dictionary file, anyway?

All in all, a thoroughly nasty experience of incomprehension and
guesswork followed by refuted hypotheses. It took me about an hour of
struggle before I hit "make uninstall".

Freeradius, OTOH, comformed to preexisting biases and convictions about
how configuration should be done quite nicely. All that I have a
problem with is the incomplete writing in the radiusd.conf file (which,
however, I *LIKE*, and I like the writing there is! It's simply
incomplete).

Peter

- 
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

Reply via email to