mo fasta and mo betta the url_for

[ahoward]

incorporated @tenderlove's suggestions.

[carlosantoniodasilva]

@ahoward great! Just a note, it's always good if you can link to the other related issue when sending a new issue / pull request, so others who didn't see it before can follow what's going on. Thanks :)

Update: related to #5743.

[ahoward]

@carlosantoniodasilva my bad. a few more:

• c41f08c
• without this central_logger makes rails 3.2.3 teh failz customink/central_logger#28 (comment)
• Issues with Rails 3.2.x ahoward/rails_default_url_options#1

[rafaelfranca]

Can we create a private method to do this?

Something like this:

def merge_hash(target, base)
  base.keys.each do |k|
    sym = k.to_sym
    target[sym] = base[k] unless target.has_key?(sym)
  end

  target
end

and call it like this:

symbolized = {}
merge_hash(symbolized, options)
merge_hash(symbolized, url_options)
_routes.url_for(symbolized)

What did you think?

[ahoward]

i did, but if you read the previous discussion people are worried about 1.0 vs 0.4 seconds for 1_000_000 iterations. adding a function call won't help you there. personally i would, but it seems contrary to the goal of making it as fast as possible. again, i'd personally go for well factored code and let the byte code writers get busy.... /cc @tenderlove

[ahoward]

@rafaelfranca this also makes the same bad assumption the current does: that target is symbolized. the bugz being caused can be summarized as

"we can't know if options or url_options has had string or symbol keys written into it"