>>
>> The concern I expressed earlier is that these types of overrides does
>> not ever belong in a proxy, and even having the possibility to enable
>> them may be bad enough. There is a very obvious conflict between the
>> two applications of Squid.
I believe override and artificial ttl/expire directives are a legitimate, and
desirable, function of a cache.
The potential is there for problems, but including the option to control all
aspects of the http transaction surely is a Good Thing? After all, it is up to
the admin to enable such hacks, and you could add a compile-time option to
include them. This should minimize the "looks-good-let's-turn-that-on"
efforts. If you want the feature, you have to go thru manpages to find it.
I think that's ample safety - even though I don't think it's OK to 'protect
users from themselves' - leave that to Microsoft and their hidden registry
settings...
As cache admins, we're not HTTP police, we're just tuning a cache to provide
maximum benefit for all. Consider 24KBps locations; my users have enough trouble
getting anywhere without some luser opening up a warez page full of popups and
banners, full of course with no-cache and immediate expiry directives...
But in the end, it's not people like myself who will be adding the features.
If I could, I would. So all we can do is try and explain how more control would
make squid more functional and useful.
btw Henrik, thanks on behalf of everyone (i spose :)) for the immense effort you
put into helping people out with their squid issues!
regards
ben
Received on Thu Nov 29 2001 - 07:35:16 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:04:43 MST