[ANSI-Smalltalk] Re: A list of interesting methods, ifNotNil:

[Richard O'Keefe]

On 25 Sep 2008, at 9:35 pm, Andres Valloud wrote:
>
> something ifNotNilDo: [:someObject | someObject doWhatever]
>
> Why is this actually useful?  I am not saying it's not, but I am
> having trouble seeing the reason.

Because the receiver, while in some sense "known", is not
as a rule *named*.

My library has

     Object >>bind: aBlock	^aBlock value: self
     Object >>bindOwn: aBlock  ^[aBlock value: self] ensure: [self  
close]

(when you are stacking stream transformations like encodings and
compression, I don't know how I would manage without #bindOwn:)
and

	<e> ifNotNil: [:x | <b>]

is equivalent to

	<e> bind: [:x | x ifNotNil: [<b>]]
>
> and skip the block argument.  One could say "ah but then this requires
> a temp var".  However, expressions such as
>
> something blah ifNotNilDo: [:something | something doWhatever]
>
> are only possible to write without parentheses when "something blah"
> does not have keyword messages in it... which is something that I try
> to avoid as much as possible...

Parentheses aren't really relevant here.
It's having a name for the result.

Me, I've always considered it better style to do

	name := expression.
	name ifNotNil: [...].