Re: includes inside squid.conf

From: Bert Driehuis <>
Date: Tue, 2 Apr 2002 01:52:47 +0200 (CEST)

On Mon, 1 Apr 2002, Alex Rousskov wrote:

> One primary reason: once you add optional_include, you will soon
> discover that what you really need are:
> optional_include_that_fails_only_if_the_file_is_not_parsable,
> optional_include_that_fails_only_if_the_file_is_not_there,
> optional_include_that_fails_only_if_the_disk_is_not_there,
> optional_include_that_fails_if_it_5pm_and_load_is_high,
> optional_include_that_fails_if_a_custom_script_fails,
> optional_include_that_never_fails_but_notifies_root_of_any_problems,
> etc.
> A simpler way is to NOT have optional includes.

Actually, I'd use an optional_include that ignored *any* issues
(including file not existing, being unparseable, etcetera). I'd love to
get rid of the external redirector I'm using and replacing it with a
file that holds acl's instead. If the semantics of this optional_include
would mean that the thing would be parsed until an error is encountered,
and the rest of the file is treated as not being there, I'd be happy.

Yes, I know my rule set had better not contain an error, but bringing
down all of my Squids simultaneously because someone added, say, a
superfluous space somewhere does not thrill me. That's why a very
limited set of people can touch our Squid configs, and we just can't do
some things we'd like to do. With optional_includes, I could delegate
the task of non-critical acl management to the help desk and be done
with it (cleaning up after them, but not having to deal with a gazillion
of support calls when Squid goes belly up).

I already have some scars on my soul for redirector stuff I botched due
to missing a newline. Failure mode counts in some situations.


                                -- Bert

Bert Driehuis, MIS -- -- +31-20-3116119
Dihydrogen Monoxide kills! Join the campaign at
Received on Mon Apr 01 2002 - 16:52:49 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:14:56 MST