Method calls and blocks and calling proc

Hey all,

I have a question about this line of code:

within “#main-menu” do
find(“h6”, :text => menu).click if menu
click_link link
end

def within(selector, &blk)
new_blk = proc do
begin
scope_selectors.push(selector)
blk.call
ensure
scope_selectors.pop
end
end
super(selector.strip, &new_blk)
end

First, when within is called, it obviously passes the string
#main-menu”, but does it also pass the block (the content between do
and end) as the second argument? The reason why I ask is because notice
within requires a second argument: &blk.

Second, we create a proc to convert a code block into an object and
store that object into new_blk local variable. Then we call the object
in the super argument list, which adds the selector (e.g. “#main-menu”)
into the scope_selectors array. Then it seems like we call the initial
code block (blk.call) adjacent to within(). But the blk code block has
not been
converted to a proc, so how is it possible to have a callable block
then?

Finally, super is only called on a parent class with a method of the
same name correct? Because in this case, within() is the only method in
entire application. super has another purpose?

Thanks for response.

On Friday, May 20, 2011 2:18:50 PM UTC-6, Ruby-Forum.com User wrote:

def within(selector, &blk)

First, when within is called, it obviously passes the string
#main-menu”, but does it also pass the block (the content between do
and end) as the second argument? The reason why I ask is because notice
within requires a second argument: &blk.

Yes it does. This is a standard feature/part of ruby. When defining a
method, you may always put, as the last entry in the list of parameters,
a
parameter with the ampersand ‘&’ prefixed to its name. This “captures”
(gives you direct access) any block passed to the method by making it
available as a Proc instance. Otherwise, any block would still be passed
in
(if you called #block_given? it would return true) but the only way to
“access” it is to call #yield.

So the #within method “captures” any block passed to it and it gets
stored
as a Proc instance, referenced by blk.

Second, we create a proc to convert a code block into an object and

store that object into new_blk local variable. Then we call the object
in the super argument list, which adds the selector (e.g. “#main-menu”)
into the scope_selectors array.

There is no “call” on the new_blk variable. The entire method definition
is
pretty much composed of the new Proc’s definition (all except the last
line
that calls #super). The “&new_blk” entry in #super’s argument list does
the
inverse of what the ampersand in #within’s method definition does.

In essence, an entry like: “&arg” in a method definition “captures” a
block
(as a proc) while a similar entry in a method call “sets it free” by
making
the proc appear just as if you’d passed a normal code block to the
method.

Example:

def asdf(&arg)
arg.call
end

Is “effectively” equivalent to

def asdf
yield
end

And then:

asdf { puts “hi” }

is effectively equivalent to:

x = proc { puts “hi” }
asdf(&x)

Then it seems like we call the initial

code block (blk.call)

Yes, inside of the new proc stored in new_blk…

adjacent to within(). But the blk code block has
not been
converted to a proc,

This was done implicitly by the #within’s method definition parameter
list’s
“&blk” entry as described above.

so how is it possible to have a callable block
then?

Finally, super is only called on a parent class with a method of the
same name correct? Because in this case, within() is the only method in
entire application. super has another purpose?

I don’t know of any other purpose. Does the “super” line result in a
NoMethodError being raised? If not, it’s finding some method up the
chain
to call. Is there some other gem or library that adds a #within method
to
Object or some other ruby magic like #method_missing somewhere up the
chain?