Reserved memcached key namespace

[Dustin Sallings]

On May 2, 2007, at 12:08 , Marc wrote:

>    quit\r\n
>
>    becomes:
>
>    set __quit__ 0 0\r\n\r\n =>
>        <= STORED\r\n
>
>    version\r\n
>
>    becomes:
>
>    get __version__\r\n =>
>        <= VALUE __version__  0 9\r\n1.2.0-rc1\r\nEND\r\n


	I like the idea for things that are accessed (version, stats,  
etc...), but I don't think it makes sense for commands that exist  
solely for their side-effects.  E.g. it doesn't help me, in authoring  
a generic interface, to have a mutation operation disconnect me.

	I'm trying to avoid using the term ``rest,'' but that's sort of what  
it feels like.  You basically need a get, delete, and put operation  
(which has three subtly different variations), and (maybe) quit.

	Flush might make sense if it uses the delete operation.  For example:

	delete __everything__ 0\r\n =>
		<= DELETED\r\n

	Note that the semantics are different between delete and flush in  
this regard, though.

	BTW, flush_regex sounds interesting, but I'd be concerned it  
couldn't be done cheaply.

-- 
Dustin Sallings


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.danga.com/pipermail/memcached/attachments/20070502/310a325c/attachment.html