[MTOS-dev] Cache::Memcached::Fast support [r1117]

[Byrne Reese]

Can you think of any other changes needed to the core to support the ability
to have a switch to turn on/off support for the driver?

And yes: MemcachedDriver makes more sense to me.



On 12/16/07 11:37 PM, "Hirotaka Ogawa" <hirotaka.ogawa at gmail.com> wrote:

> On Dec 17, 2007 3:32 PM, Byrne Reese <byrne at sixapart.com> wrote:
>> Ideally I would like to see this also support a config directive:
>> 
>> MemcacheDriver <module name>
> 
> MemcacheDriver? MemcachedDriver?
> 
>> 
>> This would default to the current mode, but alternatively one could specify
>> a separate module (included in the previous patch). We would need to
>> finalize the interface so that one could easily make these two
>> implementations swappable. This requires some docs, but I will take
>> ownership of that.
>> 
>> Then we could then include the driver and mark it as experimental until we
>> feel that it has been adequately tested by the community.
> 
> That's good.
> 
> In my patch, MT::ObjectDriver and MT::Memcached would select to employ
> C::M or C::M::Fast based on eval-test for C::M::Fast.  But, they would
> do the same thing based on a new directive.
> 
>> 
>> +1 to the contribution though (having not actually looked at the code yet) -
>> anything to help performance
>> 
>> 
>> On 12/16/07 9:05 AM, "Timothy Appnel" <tim at appnel.com> wrote:
>> 
>>> Nice work Hirotaka. The author of that module noted on the memcached
>>> mailing list that there are a few features that have not been
>>> implemented. Most notable is the stats command. (This would effect
>>> anyone using Mark Carey's plugin.) Hopefully that will be addressed
>>> shortly, but in the meanwhile we'll need to consider the tradeoff.
>>> <tim/>
>> 
>>> _______________________________________________
>>> MTOS-dev mailing list
>>> MTOS-dev at sixapart.com
>>> http://www.sixapart.com/mailman/listinfo/mtos-dev
>> 
>> 
>> 
> 
>