Would the built-in lock_version magic column be good enough for a
fragment_key?

On Oct 2, 1:09 pm, Philip Hallstrom <phi...@pjkh.com> wrote:
> > Right now, I have the :action_suffix for a fragment I cache such  
> > that I
> > can find it when I need to delete it from the cache.  But an episode  
> > on
> > Rails Casts gave the idea that another approach is to not worry about
> > deleting the old fragments.  They will eventually just get pushed  
> > out of
> > the cache.  Instead, just make sure that the name is unique to the
> > version of the object(s) for which the fragment renders.
>
> > So, the suggestion was to make the :action_suffix tied to the last
> > update field or an equivalent field.
>
> > Can anyone comment on this strategy?  It seems much easier to me.  It
> > somewhat solves the problem of trying to kill the old cache entries  
> > with
> > all the ugly sweeper logic.
>
> Seems reasonable.  As long as you're sure your cache will expire the  
> old entries and not keep piling them up until you run out of ram/disk  
> space.
>
> I would consider adding a generic method to all your models named say  
> "fragment_key" that does whatever is necessary to generate a unique  
> key.  That way in your controllers/views you never have to think what  
> fields to put in there to get the right version.  It's always just  
> @instance.fragment_key.  And you can hash that if you want to shorten  
> it down and avoid any key length issues or invalid characters, etc.
>
> You might also find this 
> interesting...http://blog.plataformatec.com.br/2009/09/how-to-avoid-dog-pile-effect...
>
> -philip
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "Ruby 
on Rails: Talk" group.
To post to this group, send email to rubyonrails-talk@googlegroups.com
To unsubscribe from this group, send email to 
rubyonrails-talk+unsubscr...@googlegroups.com
For more options, visit this group at 
http://groups.google.com/group/rubyonrails-talk?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to