Ruby undocument File.basename trick

Just stumbled upon the following trick to get the basename without an extension in Ruby:
(I didn't know you could use ".*" to remove any extension).

File.basename( "filename/sample-filename.html", ".*" )
=> "sample-filename"

Fix rails database collations

Today I discovered my Ruby on Rails development environment database migrated with the wrong character encoding. It was using the MySQL default encoding latin1.
I didn't feel throwing my database away because it contains a lot of stuff.

I used the following snippet to convert all columns to utf8 

    ActiveRecord::Base.connection.tables.each do |table|
      ActiveRecord::Base.connection.execute( "ALTER TABLE `#{table}` CONVERT TO CHARACTER SET utf8 COLLATE utf8_unicode_ci")
    end

A copy of has been removed from the module tree but is still active

I'm working on a rails website that requires me to specify a class in an initializer.

config/initializers

spree.searcher_class = MySearcherClass

I'm currently developing this searcher class. Every time I change this class I get the following message:

A copy of MySearcherClass has been removed from the module tree but is still active

This sucks big time! Because I need to restart my rails application every time I change something.

My workaround for the moment is this:

spree.searcher_class = class.new do
    def new(*args,&block)
      return MySearcherClass.new( *args, &block )
    end
  end
end

I'm not very keen on this, but it does the trick for now :)

Ruby gsub gotcha

Today I struggled with a string replace that didn't do what I expected it to do.

Consider the following code:

"xyz".gsub("y","a\\'b")
=> "xazbz"

Because gsub can be used with a regular expression the replace value can use regular expression backrefs.

I assumed (assumption is the mother of all fuckups) when using a plain string as search term, (which cannot result in back refs) it didn't use backrefs..
Well I was wrong..

A solution is to use the block-variant: 

"xyz".gsub("y") { "a\\'b" }
=> "xa\\'bz"

I think the behavior of gsub is wrong...
When you don't have a regular expression you cannot have backrefs and you can have a dumb string replace.... What's your opinion about this?