Smalltalk › Pharo › Pharo Smalltalk Users

is this valid smalltalk

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

6 messages Options

Roelof

is this valid smalltalk

Hello,

For a challenge of Exercism I need to check if some parenthes and
brackets are balanced.

so for example

() = true
([]) = true

but (]) is not true because the bracket has no opening bracket.

Now I wonder if I can premature end a #do: like this

collection do: [:element | (condition with element) ifFalse: [^false]]

in pseudo code

is this a valid way to end the do ?
or is there a better way ?

Roelof

K K Subbu

Re: is this valid smalltalk

On 04/04/19 4:46 PM, Roelof Wobben wrote:

> Hello,
>
> For a challenge of Exercism I need to check if some parenthes and
> brackets are balanced.
>
> so for example
>
> () = true
> ([]) = true
>
> but (]) is not true because the bracket has no opening bracket.
>
> Now I wonder if I can premature end a #do: like this
>
> collection do: [:element | (condition with element) ifFalse: [^false]]

Have you looked at anySatisfy:

#('the' 'quick' 'brown' 'fox') anySatisfy: [:s | s beginsWith: 'x' ]

Regards .. Subbu

Roelof

Re: is this valid smalltalk

Op 4-4-2019 om 13:22 schreef K K Subbu:

> On 04/04/19 4:46 PM, Roelof Wobben wrote:
>> Hello,
>>
>> For a challenge of Exercism I need to check if some parenthes and
>> brackets are balanced.
>>
>> so for example
>>
>> () = true
>> ([]) = true
>>
>> but (]) is not true because the bracket has no opening bracket.
>>
>> Now I wonder if I can premature end a #do: like this
>>
>> collection do: [:element | (condition with element) ifFalse: [^false]]
>
> Have you looked at anySatisfy:
>
> #('the' 'quick' 'brown' 'fox') anySatisfy: [:s | s beginsWith: 'x' ]
>
> Regards .. Subbu
>

yep, i also took a look at that one

Herby Vojčík

Re: is this valid smalltalk

In reply to this post by Roelof

On 4. 4. 2019 13:16, Roelof Wobben wrote:

Not sure I understand your question properly, but yes, you can do this,
it is called "non-local return" and is a known pattern in Smalltalk.

> or is there a better way ?

(collection
detect: [:element | (condition with element) not]
ifNone: [sentinel]) == sentinel ifFalse: [ ^false ]

for example does not force you to employ it, as well as

(collection allSatisfy: [:element | (condition with element)])
ifFalse: [ ^false ]

>
> Roelof

Herby

Richard O'Keefe

Re: is this valid smalltalk

In reply to this post by Roelof

For a change this is an exercism problem I know something about.

I did look at the exercism web site, and did all the SML exercises,

but could not find any for Smalltalk or Pharo.

The exercise is in fact about using a stack.

   stack <- empty

   for each character of the string

      if it is one of ( [ { push it on the stack

      if it is one of ) ] }

         if the stack is empty or the top element does not

            correspond to this character, return false

      if it is not one of ( ) [ ] { } just ignore it

   return (the stack is empty)

BOTH of the "return" constructs in this pseudo-code become

^ something

in Smalltalk.  "Long returns", where "^" occurs inside a block,

are well defined, very useful, and universally accepted in Smalltalk.

There is no analogue of "^" for returning from a block.

By the way, this has nothing to do with #do:.  It's about

returning from the method through any number of blocks.

For example, you will sometimes see code like

   x := aDictionary at: aKey ifAbsent: [^false].

On Fri, 5 Apr 2019 at 00:16, Roelof Wobben <[hidden email]> wrote:

Hello,

For a challenge of Exercism I need to check if some parenthes and
brackets are balanced.

so for example

() = true
([]) = true

but (]) is not true because the bracket has no opening bracket.

Now I wonder if I can premature end a #do: like this

collection do: [:element | (condition with element) ifFalse: [^false]]

in pseudo code

is this a valid way to end the do ?
or is there a better way ?

Roelof

Roelof

Re: is this valid smalltalk

I already thought to use this flow.

but I only wondered if I could use the ^false this way and appearently it is allowed.

so I can begin with this one.

The Pharo track is still hidden but I saw on the github page that in a short time it will be public with some 20 - 25 challenges.

Roelof

Op 4-4-2019 om 14:36 schreef Richard O'Keefe:

For a change this is an exercism problem I know something about.

I did look at the exercism web site, and did all the SML exercises,

but could not find any for Smalltalk or Pharo.

The exercise is in fact about using a stack.

stack <- empty

for each character of the string

if it is one of ( [ { push it on the stack

if it is one of ) ] }

if the stack is empty or the top element does not

correspond to this character, return false

if it is not one of ( ) [ ] { } just ignore it

return (the stack is empty)

BOTH of the "return" constructs in this pseudo-code become

^ something

in Smalltalk. "Long returns", where "^" occurs inside a block,

are well defined, very useful, and universally accepted in Smalltalk.

There is no analogue of "^" for returning from a block.

By the way, this has nothing to do with #do:. It's about

returning from the method through any number of blocks.

For example, you will sometimes see code like

x := aDictionary at: aKey ifAbsent: [^false].

On Fri, 5 Apr 2019 at 00:16, Roelof Wobben <[hidden email]> wrote:

Hello,

For a challenge of Exercism I need to check if some parenthes and
brackets are balanced.

so for example

() = true
([]) = true

but (]) is not true because the bracket has no opening bracket.

Now I wonder if I can premature end a #do: like this

collection do: [:element | (condition with element) ifFalse: [^false]]

in pseudo code

is this a valid way to end the do ?
or is there a better way ?

Roelof