Am I doing something with an asset pipeline?
Since the "upgrade" to Rails 3.1, my application is very slow in development mode
(> 30 per request)
I have many images, and it seems that most of this time delay is a pipeline process that processes each GET request for each image.
You do not have this problem in Staging or Production mode, as assets are cached, etc.
Is there something that I was not told, or is this how we should work now?
Requests can be slow if you have gems or parts of your application that download code at the beginning of each request, or simply link to parts of your application, resulting in most of it being downloaded. For most of them, the autoloader is the main reason for the delay in the request.
An automatic rail reloader removes any autoload classes / modules / etc at the beginning of each request and can cause significant delays at the beginning of each request, because Rails reloads all the source files it needs.
You might want to try with https://github.com/wavii/rails-dev-tweaks , which gives you detailed control over what requests force the autoloader to come in. This is really not a fix for the root cause (something does extra work at the beginning of each request, which probably shouldn't be done), but it certainly mitigates most of these problems.
If your application is slow, it is because of your application or one of the stones that you are using. I had a similar problem and it looks like Mongoid was what you can read here:
http://martinciu.com/2011/06/rails-3-1-and-slow-asset-pipeline.html