Re: Proposal: create an Extractor for PartialFunction

It would be more useful with eta-conversion, I agree.

  --Rex


On Fri, Jul 10, 2015 at 12:44 PM, 杨博 <[email protected]> wrote:

> Thank you. But I think Computes for Function1 is not that useful unless we
> allow eta-conversion as a pattern. e.g.
>
> case foo.someMethod.Computes(x) =>
>
>
>
> Rex Kerr <[email protected]>于2015年7月11日周六 上午3:32写道：
>
>> I think it's a good idea.  I'd make it more general, actually: call it
>> `Computes`, and make a (lower priority) version on Function1 also.
>>
>> You often want to do things like
>>
>> foo match {
>>   case bar.Computes(x) if baz(x) => x
>>   case _ => quux
>> }
>>
>> where the only difference with the PartialFunction is that you don't
>> always need the if clause since it's embedded.
>>
>> It wouldn't be fast for primitives, for instance, since you always have
>> the overhead of at least a virtual Some wrapper (and maybe a real one,
>> depending on the JVM).  But it'd sure be handy.
>>
>>   --Rex
>>
>>
>>
>>
>> On Thu, Jul 9, 2015 at 8:06 AM, 杨博 <[email protected]> wrote:
>>
>>>
>>> https://github.com/search?q=if+catcher.isDefinedAt&ref=reposearch&type=Code&utf8=%E2%9C%93
>>>
>>> There are too many people who have to use case e if
>>> catcher.isDefinedAt(e) =>
>>>
>>> Does that matter?
>>>
>>> 在 2015年7月9日星期四 UTC+8下午10:57:40，杨博写道：
>>>
>>>> lift and applyOrElse are hard to work with other cases in a match
>>>> expression, especially when dealing with Exceptions. For example:
>>>> https://github.com/Atry/memcontinuationed/blob/master/src/main/scala/com/dongxiguo/memcontinuationed/Memcontinuationed.scala#L599
>>>>
>>>> 在 2015年7月9日星期四 UTC+8下午10:43:15，Vlad Patryshev写道：
>>>>>
>>>>> Here's some discussion of this issue:
>>>>> http://stackoverflow.com/questions/5902266/inverse-of-partialfunctions-lift-method
>>>>>
>>>>> Instead of going back into the world of booleans and conditions, I'd
>>>>> suggest to think about why there's no such thing already available. Maybe
>>>>> this is the right direction. applyOrElse is total; if it does not
>>>>> work for you, lift should. It is pretty scientific, and gets you to a
>>>>> Kleisli category. All these ifs and match/cases are a little bit like
>>>>> crutches.
>>>>>
>>>>> Thanks,
>>>>> -Vlad
>>>>>
>>>>> On Wed, Jul 8, 2015 at 9:56 PM, 杨博 <[email protected]> wrote:
>>>>>
>>>>>> Sometimes, when I create a method, I need an extractor parameter for
>>>>>> pattern matching, while there is not a generic extractor trait in the Scala
>>>>>> Standard Library. So we have to use a PartialFunction as the
>>>>>> parameter instead.
>>>>>>
>>>>>>
>>>>>> When I use the PartialFunction, I had to create pattern matching
>>>>>> expression like this.
>>>>>>
>>>>>>
>>>>>> input match {
>>>>>>
>>>>>>   case x if partialFunction.isDefined(x) =>
>>>>>>
>>>>>>     val data = partialFunction.apply(x)
>>>>>>
>>>>>>     doSomethingWith(data)
>>>>>>
>>>>>> }
>>>>>>
>>>>>>
>>>>>> This approach is not perfect because:
>>>>>>
>>>>>>
>>>>>>    1. partialFunction.isDefined and partialFunction.apply checked x
>>>>>>    twice, which is slow if isDefined has complex logical, checking
>>>>>>    whether a file exists for example.
>>>>>>    2. The if condition around temporary variable x is too trivial.
>>>>>>
>>>>>> I think the better approach is converting the PartialFunction to an
>>>>>> extractor and matching with that extractor
>>>>>>
>>>>>>
>>>>>> implicit final class PartialFunctionAsExtractor[A, B](pf:
>>>>>> PartialFunction[A, B]) {
>>>>>>
>>>>>>   object Extractor {
>>>>>>
>>>>>>     def unapply(a: A) = PartialFunction.condOpt(a)(pf)
>>>>>>
>>>>>>   }
>>>>>>
>>>>>> }
>>>>>>
>>>>>>
>>>>>> input match {
>>>>>>   case partialFunction.Extractor(data) =>
>>>>>>     doSomethingWith(data)
>>>>>> }
>>>>>>
>>>>>> Since this is a common approach, I hope adding the
>>>>>> PartialFunctionAsExtractor to scala.Predef .
>>>>>>
>>>>>> I would like to create a pull request if people like this idea.
>>>>>>
>>>>>> --
>>>>>> You received this message because you are subscribed to the Google
>>>>>> Groups "scala-language" group.
>>>>>> To unsubscribe from this group and stop receiving emails from it,
>>>>>> send an email to [email protected].
>>>>>> For more options, visit https://groups.google.com/d/optout.
>>>>>>
>>>>>
>>>>>  --
>>> You received this message because you are subscribed to the Google
>>> Groups "scala-language" group.
>>>
>> To unsubscribe from this group and stop receiving emails from it, send an
>>> email to [email protected].
>>>
>>
>>> For more options, visit https://groups.google.com/d/optout.
>>>
>>  --
>> You received this message because you are subscribed to a topic in the
>> Google Groups "scala-language" group.
>> To unsubscribe from this topic, visit
>> https://groups.google.com/d/topic/scala-language/g0-hbN5qerQ/unsubscribe.
>> To unsubscribe from this group and all its topics, send an email to
>> [email protected].
>> For more options, visit https://groups.google.com/d/optout.
>>
> --
> 杨博 (Yang Bo)
>
> --
> You received this message because you are subscribed to the Google Groups
> "scala-language" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> For more options, visit https://groups.google.com/d/optout.
>

-- 
You received this message because you are subscribed to the Google Groups "scala-language" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
For more options, visit https://groups.google.com/d/optout.