[swift-evolution] [swift-evolution-announce] [Review] SE-0155: Normalize Enum Case Representation

Matthew Johnson matthew at anandabits.com
Sat Feb 25 08:13:24 CST 2017 


Sent from my iPad

> On Feb 24, 2017, at 11:26 PM, Daniel Duan via swift-evolution <swift-evolution at swift.org> wrote:
> 
> Before I start revising this proposal, there are a couple of open questions I’d like to discuss with the community and the core team.
> 
> The first question relates to the purpose of having a “internal” argument name. There are applications of such names in GADT (if we ever get there) and perhaps the case-as-subtype-of-the-enum stories on the list right now. Out side of these scenarios, however, such names has few chances to be used. The one I can come up with, which is also the “open” part of the question, is this: we can use the internal names in pattern matching, as opposed to using the labels. This seems to align with the subtyping/GADT use cases. Is this a desirable outcome?

This is what I suggested.  I think of it like this: the case includes a static factory method that produces values.  We obviously want the external label here.  The pattern is like matching a labeled tuple.  When we think of the cases as a labeled tuple (or synthesized struct in the case-as-subtype world) it is clear (to me at least) that we want to use the internal name.  To avoid the noise, I suggest allowing patterns to elide the label if they bind to an identical name, which is often what they will do.

> 
> The second open question is the syntax for “overloaded” cases. If we decide to allow them, what should the patterns matching them look like? I can think of one obvious-ish design where we make the pattern look like the declaration and require types for disambiguation. So the most verbose form of pattern would look something like
> 
> ```
> case let .baseName(label0 name0: Type0, label1 name1: Type1)
> ```

We don't need new syntax for this.  Cast patterns in the associated value slots should work just fine:

case let .baseName(name0 as Type0, name1 as Type1)

Here, `name0` and `name1` are the internal names of the pattern.  If the pattern had different internal names the label must be used:

case let .baseName(internalName1: name0 as Type0, internalName2: name1 as Type1)

Note: the syntactic sugar of eliding the label should also be allowed in cases that don't don't distinguish the internal and external names.

> 
> Even in this design, there are the choice of forcing name0/name1 to match the “internal” name in declaration and letting users choose whatever they want. There maybe way elegant syntax that I haven’t thought of. Since my experience with languages with pattern matching syntax is pretty much limited to SML and Haskell, I must admit my inadequacy here. Any help is appreciated here!
> 
> 
>>> On Feb 22, 2017, at 11:27 AM, John McCall via swift-evolution <swift-evolution at swift.org> wrote:
>>> 
>>> On Feb 21, 2017, at 4:43 AM, Daniel Duan via swift-evolution <swift-evolution at swift.org> wrote:
>>> My apologies for misunderstanding.
>>> 
>>> Would it be better to add the anonymous case feature in a separate proposal? It stands alone as a new addition to enum. The intent for this proposal is bringing enum's syntax closure to other part of Swift.
>> 
>> The core team met and talked about SE-0155 today, even though we're not quite done with the review period, and here's where we stand.
>> 
>> SE-0155 is being returned for revision.  Consensus appears to be strongly in favor of requiring argument labels (if provided) to be enforced on "call" sites.  However, the core team feels that the proposal needs revision in the following ways:
>>   - Are internal argument names syntactically allowed in a case declaration?
>>   - Can cases with the same base name be overloaded by argument label?  If so, is a pattern match using just the bare name ambiguous or does it match both cases?
>>   - Can cases with the same name (using the same rule as above) be overloaded by argument type?  If so, how are they disambiguated in pattern matches?
>>   - Do pattern matches require argument labels to be given along with value patterns, e.g. "case .valid(value: let value)", or is there some way to shorten this?  If the latter, what are the rules for that?
>>   - Are you proposing anonymous cases, and if so, what are the language rules for them?
>>   - The proposal makes a claim about layout efficiency; please either discuss this claim or remove it.
>> 
>> John.
>> 
>>> 
>>> Daniel Duan
>>> Sent from my iPhone
>>> 
>>>> On Feb 21, 2017, at 1:15 AM, Dave Abrahams via swift-evolution <swift-evolution at swift.org> wrote:
>>>> 
>>>> I had not intended for _ to be an ordinary identifier, but as a way of spelling the empty case name. Ambiguous cases, not distinguished by either full name or payload type, would be errors. How to construct anonymous cases?  I guess it would be MyEnum(expression)
>>>> 
>>>> Sent from my moss-covered three-handled family gradunza
>>>> 
>>>>> On Feb 19, 2017, at 2:52 PM, Daniel Duan <daniel at duan.org> wrote:
>>>>> 
>>>>> 
>>>>>>> On Feb 19, 2017, at 11:49 AM, Matthew Johnson via swift-evolution <swift-evolution at swift.org> wrote:
>>>>>>> 
>>>>>>> 
>>>>>>> On Feb 18, 2017, at 10:49 PM, Dave Abrahams via swift-evolution <swift-evolution at swift.org> wrote:
>>>>>>> 
>>>>>>> I'm on vacation and don't have time for a full review right now, but I am concerned that wild this proposal would make enums more general and uniform with the rest of the language , they also would become much more awkward for common use cases. I have recently been very pleased that I didn't have to supply labels in switch statements where the label name would simply have matched the name of the variable to be bound.  This looks needlessly verbose:
>>>>>>> 
>>>>>>>   case .valid(value: let value, resumptionPoint: let resumptionPoint):
>>>>>>> 
>>>>>>> I cannot imagine a real life use case where one would have labels in the case and desire to bind associated values to variables having different names than the labels.
>>>>>> 
>>>>>> I agree with this, but I think it’s an issue we can solve (perhaps as an amendment to this proposal).
>>>>>> 
>>>>>> First, I think Brent’s idea of introducing an argument label that can be distinct from the “property” name of the case is a good one.  I think we should do this.  It takes the parallel with function signatures even further.
>>>>>> 
>>>>>> Second, we should allow the “property” name to be `_`.  This would mean no label can be used when matching:
>>>>>> 
>>>>>> case valid(value _: ValueType, resumptionPoint _: PointType)
>>>>>> 
>>>>>> Third, I think we should also allow suers to elide the label if they either discard the value with `_` or bind a name that is identical to the label, so we might have:
>>>>>> 
>>>>>> // declaration:
>>>>>> case valid(externalCasConstructorLabel value: ValueType, externalCaseConstructorLabel resumptionPoint: PointType)
>>>>>> 
>>>>>> // match ok:
>>>>>> case .valid(let value, let resumptionPoint):
>>>>>> 
>>>>>> // error, names do not match:
>>>>>> case .valid(let foo, let bar):
>>>>>> 
>>>>>> // ok, label is used:
>>>>>> case .valid(value: let foo, resumptionPoint: let bar):
>>>>>> 
>>>>>> This follows the behavior of function signatures very closely.  The external label is used to provide context for the argument at the call site (of the case constructor).  The internal name is used to bind a name to the value that is used by code that works with the value.  
>>>>>> 
>>>>>> The only exception here is that because the usage site is distant from the case declaration it may wish to use a different name.  We allow that, but only if the “internal name” is also used in the pattern.  This preserves the ability of a reader of the code to see the name / meaning of the associated value as it was declared by the enum in addition to the name that might make more sense for use in the local context.
>>>>>> 
>>>>>>> 
>>>>>>> Secondly, I can't imagine a case where one would want to use the same case basename and different labels. The very common use case where the types of associated values completely distinguish the case and one would rather not have to supply a case name at all is completely unaddressed. If my quick read is not mistaken, this proposal makes it legal for cases to have different complete names (including base name and labels), but doesn't make it legal to have the same full name (which I would love to be "_" or missing in some cases) with different associated value types. If we were truly following the precedent set by function signatures, wouldn't that be possible too?
>>>>>> 
>>>>>> +1.  I think this makes a lot of sense.  It completes the parallel of cases with overloaded functions.
>>>>>> 
>>>>>> I think anonymous cases are a really good idea.  I discuss those quite a bit in the value subtyping manifesto I shared last week (I’d love to hear your thoughts on it if / when you have time to take a look).
>>>>>> 
>>>>>> How would you propose that values of anonymous cases be constructed and matched?  My solution is to allow them to be constructed by implicit conversion from the associated value type to the enum type and matched by a cast pattern.  Is that what you have in mind?  I would *really* love to see this someday...
>>>>> 
>>>>> I can’t speak for Dave obviously. But I think he was merely proposing “overloaded” form of enum options, in which multiple options may share the compound name but with differently associated types. The name “_” would just be a normal identifier in such scenario. So it would also be the contractor’s function name.
>>>>> 
>>>>>>> 
>>>>>>> Sent from my moss-covered three-handled family gradunza
>>>>>>> 
>>>>>>>> On Feb 17, 2017, at 5:26 PM, John McCall <rjmccall at apple.com> wrote:
>>>>>>>> 
>>>>>>>> Hello Swift community,
>>>>>>>> 
>>>>>>>> The review of "SE-0155: Normalize Enum Case Representation" begins now and runs through next Friday, February 26th. The proposal is available here:
>>>>>>>> 	https://github.com/apple/swift-evolution/blob/master/proposals/0155-normalize-enum-case-representation.md
>>>>>>>> 
>>>>>>>> Reviews are an important part of the Swift evolution process. All reviews should be sent to the swift-evolution mailing list at
>>>>>>>> 	https://lists.swift.org/mailman/listinfo/swift-evolution
>>>>>>>> or, if you would like to keep your feedback private, directly to the review manager. When replying, please try to keep the proposal link at the top of the message:
>>>>>>>> 
>>>>>>>> 	Proposal link: https://github.com/apple/swift-evolution/blob/master/proposals/0155-normalize-enum-case-representation.md
>>>>>>>> 
>>>>>>>> 	Reply text
>>>>>>>> 
>>>>>>>> 	Other replies
>>>>>>>> 
>>>>>>>> What goes into a review?
>>>>>>>> 
>>>>>>>> The goal of the review process is to improve the proposal under review through constructive criticism and, eventually, determine the direction of Swift. When writing your review, here are some questions you might want to answer in your review:
>>>>>>>> 
>>>>>>>> 	• What is your evaluation of the proposal?
>>>>>>>> 	• Is the problem being addressed significant enough to warrant a change to Swift?
>>>>>>>> 	• Does this proposal fit well with the feel and direction of Swift?
>>>>>>>> 	• If you have used other languages or libraries with a similar feature, how do you feel that this proposal compares to those?
>>>>>>>> 	• How much effort did you put into your review? A glance, a quick reading, or an in-depth study?
>>>>>>>> 
>>>>>>>> More information about the Swift evolution process is available at https://github.com/apple/swift-evolution/blob/master/process.md
>>>>>>>> 
>>>>>>>> Thank you,
>>>>>>>> 
>>>>>>>> John McCall
>>>>>>>> Review Manager
>>>>>>>> _______________________________________________
>>>>>>>> swift-evolution-announce mailing list
>>>>>>>> swift-evolution-announce at swift.org
>>>>>>>> https://lists.swift.org/mailman/listinfo/swift-evolution-announce
>>>>>>> _______________________________________________
>>>>>>> swift-evolution mailing list
>>>>>>> swift-evolution at swift.org
>>>>>>> https://lists.swift.org/mailman/listinfo/swift-evolution
>>>>>> 
>>>>>> _______________________________________________
>>>>>> swift-evolution mailing list
>>>>>> swift-evolution at swift.org
>>>>>> https://lists.swift.org/mailman/listinfo/swift-evolution
>>>>> 
>>>> _______________________________________________
>>>> swift-evolution mailing list
>>>> swift-evolution at swift.org
>>>> https://lists.swift.org/mailman/listinfo/swift-evolution
>>> _______________________________________________
>>> swift-evolution mailing list
>>> swift-evolution at swift.org
>>> https://lists.swift.org/mailman/listinfo/swift-evolution
>> 
>> _______________________________________________
>> swift-evolution mailing list
>> swift-evolution at swift.org
>> https://lists.swift.org/mailman/listinfo/swift-evolution
> 
> _______________________________________________
> swift-evolution mailing list
> swift-evolution at swift.org
> https://lists.swift.org/mailman/listinfo/swift-evolution
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.swift.org/pipermail/swift-evolution/attachments/20170225/3bc10762/attachment.html>

More information about the swift-evolution mailing list