Jeff Dean

Sorry - that patch didn't come through too well on 2-3-stable. Here's an updated version for 2-3-stable. It applies cleanly for me, and tests pass.

Thank you for the feedback.

I'll find a good way to remove that loop and resubmit.

This issue is far more complex than I realized, but luckily the fix was simple and clean. Take this example:

class Client < Company
  belongs_to :firm, :foreig...

The first method ignores that id in all database drivers. first just adds a limit 1 to the query - so in your case, it looks like in your mysql database, the fi...

I agree that the db errors should be consistent, and I believe they are. However, when you execute remove_column(:name) it's never getting to mysql at all becau...

Thank you for verifying.

The block syntax of fetch only evaluates the default value if the key is not present in the hash.

In the options.fetch(:foo, bar) case,...

The attached patch contains:

• a fix for the nil-id
• tests
• a minor refactoring to the fetch syntax for better performance (using the block syntax instead of the a...

Currently you cannot set the value of a text_field or hidden_field to nil:

f.text_field :some_field, :value => nil # => will output an input with a value equal...

Pratik - any chance you could elaborate on why routes should be case-sensitive?

Are you aware of any apps in production that rely on case-sensitive routes? Is t...

This has happened to me in the past when loading a gem or plugin that requires rake. For example whenever.

If you are using bundler, you may have to add :requir...