Improving Compiler Error Messages

[Jacob Carlborg]

On 5/4/10 00:30, Michel Fortin wrote:
> On 2010-05-03 18:05:11 -0400, Walter Bright <newshound1 at digitalmars.com>
> said:
>
>> Nick Sabalausky wrote:
>>> "Walter Bright" <newshound1 at digitalmars.com> wrote in message
>>> news:hrncpk$14t2$2 at digitalmars.com...
>>>> Michel Fortin wrote:
>>>>> That's an argument against returning the value of the last
>>>>> statement. It's not an argument about making the last semicolon
>>>>> optional, is it?
>>>> Yes, it is an argument against it. It means the semantics can vary
>>>> dramatically depending on whether a ; is present or not.
>>>
>>> But that's only true when the language, unlike D, has implicit
>>> returns, right?
>>
>> Right. But D (and C, C++) do not have implicit returns, and so this
>> syntax is not right for them.
>
> Ok, so to summarize, your opinion is:
>
> 1. omitting the last semicolon looks like an implicit return
> 2. implicit returns are too subtle, won't add to D
> 3. since D has no implicit returns, omitting the last semicolon is not
> accepted
>
> I disagree about 1, and thus about 3 since it derives from 1. But that's
> probably just because of our different background, so I'll leave it at
> that.

I guess D isn't created with delegate literals in mind as one of the 
most important features, like Ruby or Scala.