Re: [squid-users] squid-3 problem with "cache_dir null"

From: Amos Jeffries <squid3@dont-contact.us>
Date: Tue, 18 Dec 2007 23:57:28 +1300

Adrian Chadd wrote:
> On Tue, Dec 18, 2007, Amos Jeffries wrote:
>
>>> Yeah, but what it doesn't make abundantly clear is that the path isn't
>>> an option, its required for all storedir types. :)
>> It really should not be for null.
>> null should have an implicit and required "/dev/null" in the source code
>> if dir string is required for anything like display or hashes.
>>
>> I assume its still in there for netdb and maybe swap.state?
>
> Its there for legacy reasons.
>
> Ideally, nothing "special" should be using the first storedir, like netdb.
> (And that fails anyway if you use / and run Squid as non-root, so it should
> be changed :)
>
> The swap.state isn't used for the null fs, its never opened, closed or checked.
> Its a holdover from way long ago, back in "modio" branch days, when I broke
> the storage code out into modules. Every storage module back then had a
> path, so I put "path" as being something in the SwapDir struct and parsed before
> options were handed off to the per-storedir type parser. Another example of
> an annoyingly-chosen line in the sand.

Which means once the netdb is fixed as per Henriks comments earlier it
can dropped from cache_dir null in next round of squid.conf cleanups.

Thank you.

Amos

-- 
Please use Squid 2.6STABLE17 or 3.0STABLE1.
There are serious security advisories out on all earlier releases.
Received on Tue Dec 18 2007 - 03:57:24 MST

This archive was generated by hypermail pre-2.1.9 : Tue Jan 01 2008 - 12:00:02 MST