Take Hash with options inside Array in #url_for

[ognevsky]

Allows you to code like this:

= link_to 'Edit', [:edit, @post, { author_id: current_user.id }]

[ognevsky]

cc @drogus

[carlosantoniodasilva]

Needs tests and a changelog entry.

[senny]

great patch 👍 looking forward to it!

[senny]

/cc @bjoernbur

[bjoernbur]

thanks for this patch...

[ognevsky]

I'm done with tests and a changelog entry.

cc @carlosantoniodasilva @senny

[carlosantoniodasilva]

Please use 1.9 style hash.

[ognevsky]

Done.

cc @carlosantoniodasilva

[egilburg]

+1 for the behaviour, but why extract_options! is needed for the Array case but not for the basic else case? Could perhaps the fix better live inside the polymorphic_path/polymorphic_url methods themselves?

[ognevsky]

@egilburg there are a lot of else cases, not only arrays. So, it would break a lot of tests and code (actually, I tried it in my first attempt).

Look, polymorphic helpers already have this functionality, you can path url options as the second param (as I do with extract_options!).

You can use it like polymorphic_path [:some, :path, :here], param: 'value', the problem is that you can't set this param without polymorphic_path invocation, so you can't actually write link_to post.title, [:edit, :admin, @post], param: 'value'.

I hope I've cleared this.

[pixeltrix]

@egilburg doing it in polymorphic_path was tried in #7259 and rejected because of clashes with form_for behaviour.

@josevalim are you still 👎 on this behavior ?

[senny]

@josevalim ping

[ognevsky]

@senny ping

[senny]

@ognevsky we need to wait for @josevalim to jump in.

[ognevsky]

@senny ok, let's wait.

[josevalim]

The issue here is what is going to be considered the form object in form_for if the last element is a hash. Can a hash be a form object today? If so, how this patch affect form_for? If we have those semantics clear, I don't see any problem with this patch.

[pixeltrix]

@ognevsky if you can add tests for form_for with this behaviour and rebase so it'll merge cleanly then I think we have a 👍

[ognevsky]

@josevalim @pixeltrix sorry for the long wait.

Jose, this patch doesn't affect to form_for, because it uses polymorphic_path here: https://github.com/rails/rails/blob/master/actionview/lib/action_view/helpers/form_helper.rb#L453, not the url_for.

So, I don't change form_for behavior.

[ognevsky]

@josevalim @pixeltrix I've rebased and updated this request.
I'll add a record to changelog before the merge, it changes very often.

[chancancode]

@ognevsky Looks like there's failure https://travis-ci.org/rails/rails/jobs/13876911#L3127

[ognevsky]

@chancancode oops, I'll fix it in the morning

[ognevsky]

Fixed. I've added this failing spec to the commit accidentally, it should never pass (because form_for doesn't have this features)

[pixeltrix]

@ognevsky can you add a CHANGELOG entry and squash your commits - then I'll merge it.

[ognevsky]

@pixeltrix I'm ready

[pixeltrix]

@ognevsky thanks!

[nathanbertram]

You da man!