acts_as_cached Google Group

Re: Failing specs on Rails 2.1

n...@wundrbar.com (ngrandy) — Mon, 30 Jun 2008 18:47:57 UT

my fragment caching also broke when i upgraded to rails 2.1. rather
than debug it, i actually switched over to using the fragment caching
that is built into rails. 2.1 has improved caching a lot so that for
fragment caching it basically serves the same as cache_fu. there is
one patch here that did not make it into 2.1 - this patch allows

Failing specs on Rails 2.1

elijah.mil...@gmail.com (Eli Miller) — Mon, 30 Jun 2008 01:45:46 UT

I was about to bring some of my patches up to current code and I am
getting a lot of failing specs. Are these failing for everyone or just
me?
I have the latest version of gem Rails and cache_fu from github.
17 tests, 2 assertions, 1 failures, 16 errors
It looks like some 2.1 cache changes, here are the two distinct errors:

Re: [acts_as_cached] after_save :expire_cache pattern can lock up your database transaction

elijah.mil...@gmail.com (Eli Miller) — Mon, 30 Jun 2008 01:39:46 UT

I've been using after_commit successfully on my project for a long
while now. Do a google search or check the mailing list archives. It
has some faults I've been meaning to address, such as:
1. it always runs, even if the save fails
Easy optimization.
2. it can still run inside of another user-called transaction.

Re: Equivalent of reset_cache for fragments?

lee.fy...@gmail.com (Lee Fyock) — Fri, 27 Jun 2008 18:11:52 UT

Have you found any good approaches to this?

I was thinking of implementing something in fragment_cache.rb that is
similar to what I did for regular cache sets/gets: <http://
groups.google.com/group/acts_a s_cached/browse_thread/thread/
fc81947e9e0eb0c3>.

That is, if a cached fragment is close to expiring, set a sentinel for

Re: after_save :expire_cache pattern can lock up your database transaction

lee.fy...@gmail.com (Lee Fyock) — Fri, 27 Jun 2008 18:03:33 UT

Our memcache servers are on the same backplane as the application and
database servers, so I've never run into a situation where the
connection to memcache takes a while.

I do like the idea of after_commit, though. We use acts_as_solr, and
it occasionally puts something into the solr index that doesn't

after_save :expire_cache pattern can lock up your database transaction

bke...@gmail.com (Brandon Keene) — Sun, 22 Jun 2008 04:45:06 UT

Hey all,
This is somewhat of a question and a warning.

So I have some seemingly innocuous cache_fu code, like so

class Business
acts_as_cached :include => :reviews

after_save :expire_cache
end

Seems sane, but it has caused us some major headaches.

So ActiveRecord wraps destructive database actions in a transaction.

Re: [acts_as_cached] Re: Conditional action-caching stops after a period of time

doug.sparl...@gmail.com (Doug Sparling) — Thu, 12 Jun 2008 19:15:42 UT

I was in the middle of a site launch and this came up - Jan fortunately
helped me and it worked and I haven't had time to revisit it - I had meant
to post that back to the list a month ago.
==
doug

Re: [acts_as_cached] Re: Conditional action-caching stops after a period of time

codesu...@gmail.com (Sujal Shah) — Thu, 12 Jun 2008 17:43:18 UT

Yeah, this is one way to do it. Might be better than my solution.
Basically, he was trying to keep the app from writing out the current
rendered view to the cache if the conditional was false. Deleting it
from @actions did it, but it also deleted it for any future calls (the
filter is instantiated once and kept around rather than being created

Re: Conditional action-caching stops after a period of time

doug.sparl...@gmail.com (dsparling) — Thu, 12 Jun 2008 15:08:42 UT

Sujal,

It is indeed something with that line you posted. I contacted the
author of the original post (Jan Prill) and he sent me his patch,
which I've been meaning to post here. I had to modify it slightly for
Rails 2 and I'm glad to say that cache_fu has been working great ever
since.

/vendor/plugins/cache_fu/lib/a cts_as_cached/fragment_cache.r b

Re: [acts_as_cached] Re: Conditional action-caching stops after a period of time

codesu...@gmail.com (Sujal Shah) — Wed, 11 Jun 2008 02:29:08 UT

spariam,
I think you're running into the issue I just described in my previous
email to the list. I figured out what the line I asked about is for,
and I'm working out a patch now. Hopefully, this will solve your
problem, too.
Sujal

Fragment Cache conditionals

codesu...@gmail.com (Sujal Shah) — Wed, 11 Jun 2008 02:09:41 UT

I'm confused by a line in the fragment_cache.rb code and was hoping
someone could shed some light on it for me.
Line 74 in fragment_cache.rb reads like so:
@actions.delete(controller.act ion_name.intern) if conditional == false
I don't really get what that's trying to do. As far as I can tell,
and this is showing up in my testing, this code ends up causing the

Equivalent of reset_cache for fragments?

l...@lukemelia.com (Luke Melia) — Tue, 03 Jun 2008 08:10:42 UT

I'm putting in some fragment caching and implemented some sweepers to
expire the fragments. So far, so good. But I'd really like to freshen
up the cached fragments without them every going stale and being
cached by users.

I was thinking of having a background queue that basically does a curl
with a flag that skips the cache get and performs the cache write. Is

Associations

brandon.ca...@gmail.com (Brandon) — Sat, 31 May 2008 22:06:02 UT

cache_fu's capabilities with Ar associations are not clear to me.

Adding acts_as_cached :include => :association_name to a model appears
to do nothing, or maybe it's just not what I expect. Here are my
models.

class Channel < ActiveRecord::Base
has_many :media_mounts

after_save :reset_cache

Caching has_many :through And Eviction

cwdi...@gmail.com (s.ross) — Tue, 27 May 2008 21:13:21 UT

Hi--

I have a question that's probably been asked and answered, but Google
(the source of all knowledge :) hasn't turned up anything I feel
confident with. I have a hm:t relationship and am concerned that the
join table could be cached, but objects in the joined table could be
evicted:

This site is searched often, and many searches are identical. The

fragment caching

wojtek.ogrodowc...@gmail.com (Wojtek Ogrodowczyk) — Fri, 23 May 2008 11:19:04 UT

I'm using cache_fu for session, custom finders and fragment caching.
The first two work as expected, but all the fragment caching happens
via filesystem (the default Rails way).

I have something like this in my views:
<% cache( "/users/#{if logged_in? then current_user.id else 'false'
end}/show/home_greet", { :ttl => 1.hour }) do %> some html