Rails 3.1, Sprockets and Amazon S3?

Hi everyone,

I am working on my first Rails 3.1 app and would like to serve
production
assets from s3. In 3.0 I use jammit-s3. What’s the recommended way of
doing
this in 3.1, especially when thinking about asset versioning through
path
instead of query parameter? I vaguely remember that this was addressed
in a
slide or video but I can’t track it down.

Thanks
Martin

The config keys I’ve found are pertinent are:
config.assets.enabled = true
and
config.action_controller.asset_host

From what I can tell, the basic functionality is there, although it
might
require a (either custom or standardized) rake task or the such to
accomplish the actual push to your s3 bucket.

One of the issues I’ve been facing is the when you use the
sprockets-based
asset pipeline, it likes to add a cache-bursting hash onto the file’s
name,
which isn’t included when using the rail’s asset_path view helper
(even
the sprocket’s version).

Would seem to me that this kind of functionality should be
‘end-user/developer’ accessible but where to put the functionality: in
sprockets or on the rails side?

Hi Colin,

thanks for the reply. I found this link:
http://code.dblock.org/ShowPost.aspx?Id=205 which is quite helpful, and
jammit-s3 includes some code that could be copied as well.

To answer your question, I think it probably should be put into
sprockets,
as sprockets is handling all the name mangling and image referencing,
and
other tools like jammit-s3 have different strategies for versioning
assets.