[OpenSER-Devel] SF.net SVN: openser: [3163] trunk/modules/permissions/mi.c

[Dan Pascu]

I wonder if we should send a 200 OK for this case. Since the database 
reload is not necessary because caching is disabled, having this 
operation be a NO-OP is tempting to allow a provisioning frontend to 
issue a reload command that works on both a caching-enabled proxy as well 
as on a caching-disabled one.

500 is a too generic error which makes this case hard to distinguish and 
in the end sending a trusted reload to a proxy that doesn't cache the 
trusted table should not result in an error but at most in a warning IMO

If people are not comfortable with sending a 200 OK in this case because 
it masks a potential issue, maybe we should use another 2xx code to 
indicate that the operation was still successful, but there is a 
difference compared to a normal table reload.

Anyway, I thing this should either be considered a successful operation or 
have a result code that is easily discernable, preferable a positive one. 
It should not have the generic 500 error code at all.

On Friday 16 November 2007, Bogdan-Andrei Iancu wrote:
> Revision: 3163
>           http://openser.svn.sourceforge.net/openser/?rev=3163&view=rev
> Author:   bogdan_iancu
> Date:     2007-11-16 08:03:58 -0800 (Fri, 16 Nov 2007)
>
> Log Message:
> -----------
> - fixed trusted_reload MI command - disable it if cache is not used.
>   Reported by Dan Pascu.
>
> Modified Paths:
> --------------
>     trunk/modules/permissions/mi.c

-- 
Dan