cc'ing the list back in to catch up on this.

-- E

> On Mar 16, 2016, at 10:54 AM, Chris Wood <[email protected]> wrote:
> 
> I really need to remember the “reply to all” button for these mailing lists. 
> *Facepalm* :)
> 
> Chris
> 
>> On Mar 16, 2016, at 4:25 PM, Chris Wood <[email protected]> wrote:
>> 
>> I suspect it does (haven’t tested it), but the chances of there being a 
>> function that returns Void and a line of code that returns Void are fairly 
>> good (especially in UI code).
>> 
>> Chris
>> 
>>> On Mar 16, 2016, at 4:16 PM, Erica Sadun <[email protected]> wrote:
>>> 
>>> Well that makes sense to me honestly unless you think there should be some 
>>> condition that 
>>> a return value should at least start on the same line as the return 
>>> keyword. Although you'd
>>> think the closure would know if it's -> Void or -> T and parse the return 
>>> accordingly
>>> 
>>> -- E
>>> 
>>>> On Mar 16, 2016, at 9:37 AM, Chris Wood <[email protected] 
>>>> <mailto:[email protected]>> wrote:
>>>> 
>>>> Erica:
>>>> 
>>>> The return case is (supposedly) handled as a compiler warning. It does 
>>>> warn that code on following lines will be treated as an argument to the 
>>>> return call normally (it does in a simple test project or playground). 
>>>> Except in the code I’m working on, it didn’t warn and happily compiled 
>>>> (and there was much swearing as a result!)
>>>> 
>>>> That looks like a compiler bug, and I’ve filed a radar (25192157) already.
>>>> 
>>>> Personally I think the compiler warning (when it works) and a trailing 
>>>> semicolon is acceptable in this case (as there may be a time when I want 
>>>> arguments to ‘return’ on the next line for some obscure reason).
>>>> 
>>>> Chris Wood
>>>> 
>>>> 
>>>>> On Mar 16, 2016, at 3:30 PM, Erica Sadun <[email protected] 
>>>>> <mailto:[email protected]>> wrote:
>>>>> 
>>>>> -1
>>>>> 
>>>>> I think end-of-line semicolons are ugly. I also think the decision to use 
>>>>> them or not use them
>>>>> should be left to individuals and not enforced by the compiler.
>>>>> 
>>>>> In the current implementation a semicolon can be optionally added after 
>>>>> any statement and
>>>>> is proactively used to separate statements on a single line. Changing the 
>>>>> language to enforce
>>>>> what appears to be a style choice is not something I'd support.
>>>>> 
>>>>> The issue with the return needing a semicolon sounds odd to me. I'm 
>>>>> curious as to whether
>>>>> that's something that needs fixing.
>>>>> 
>>>>> -- E
>>>>> 
>>>>> On Mar 16, 2016, at 9:25 AM, Chris Wood via swift-evolution 
>>>>> <[email protected] <mailto:[email protected]>> wrote:
>>>>> 
>>>>> A word of warning on this - I’ve just hit a case where an end of line 
>>>>> semicolon is actually critical!
>>>>> 
>>>>> If you’re debugging and need an early exit from a function, you add 
>>>>> “return” somewhere in the middle of the code. But return can take an 
>>>>> argument, so the compiler takes the argument from the following line 
>>>>> which you think won’t be executed. Bad things happen, followed by 
>>>>> swearing.
>>>>> 
>>>>> So in that particular case, you do in fact need to write:
>>>>> 
>>>>> return;
>>>>> 
>>>>> There may be other edge cases where it’s necessary to explicitly mark the 
>>>>> end of something in the same way.
>>>>> 
>>>>> Chris Wood
>>>>> 
>>>>> 
>>>>> Hi,
>>>>> 
>>>>> I submitted a PR with a proposal to remove the swift end of line 
>>>>> semicolons.
>>>>> 
>>>>> It was rejected because i didn't discuss it here. So here i'm discussing 
>>>>> it :)
>>>>> 
>>>>> My proposal is simple: remove the semicolons in the end of lines.
>>>>> It isn't needed and makes the code ugly.
>>>>> It must be decided wether to use it or not for every project we start in 
>>>>> the coding style.
>>>>> 
>>>>> What do you think?
>>>>> 
>>>>> João Nunes
>>>>> 
>>>>> 
>>>>> Sent from my iPhone
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> _______________________________________________
>>>>> swift-evolution mailing list
>>>>> [email protected] <mailto:[email protected]>
>>>>> https://lists.swift.org/mailman/listinfo/swift-evolution

_______________________________________________
swift-evolution mailing list
[email protected]
https://lists.swift.org/mailman/listinfo/swift-evolution

Reply via email to