--
Cheers

Koz

On 27/09/2012, at 7:39 PM, Xavier Noria <[email protected]> wrote:

Let me propose something.

Since we are already using Entry, Brian's approach is written, and it is a
drop-in bringing great improvement for many use cases over the current
store (caching counters, flags, small strings, etc.), I'd suggest that we
work on a pull request with that patch. This option is very concrete and
doable in the short term. That way we make sure Rails 4 ships with *at
least* these improvements.

Then, if someone can go further and make dalli report missing keys somehow,
or want to write a plugin which does not use Entry at the cost of not
supporting nil and race_condition_ttl, perfect. And if the former happens,
then let's see what do we do.

Any objection?


Yes, but only on the basis that relying on this behavior is a truly bad
idea if you're actually using memcached at a large scale. But clearly there
are other views on this functionality.

Given that clearly some people want this then using Entry should be an
optional extra which defaults to off. Or perhaps ship another memcached
store subclass with the bells and whistles.

That let's people use the race condition ttl stuff and the cache-nil stuff
if they need it, but doesn't apply a non-trivial per-key overhead to
everyone who uses memcached.

 --
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Core" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/rubyonrails-core?hl=en.

-- 
You received this message because you are subscribed to the Google Groups "Ruby 
on Rails: Core" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/rubyonrails-core?hl=en.

Reply via email to