Biggest Issue with D - Definition and Versioning

[Alex Rønne Petersen]

On 12-01-2012 08:30, Brad Roberts wrote:
> On 1/11/2012 11:16 PM, Jacob Carlborg wrote:
>> On 2012-01-11 21:34, Walter Bright wrote:
>>> On 1/11/2012 12:29 PM, Zachary Lund wrote:
>>>> D1 or D2 both don't have a very accurate definition. C and C++ has
>>>> several
>>>> hundred pages of specifications and while I'm hopefully not expecting
>>>> that, I do
>>>> want some certainty on what is *supposed* to happen.
>>>>
>>>> There is also no versioning. D2 is definitely different from D1 in the
>>>> case that
>>>> it's not backwards compatible, but version for the actual example
>>>> implementation
>>>> isn't enough, especially when that implementation doesn't always do
>>>> what it's
>>>> supposed to do.
>>>
>>> Please file bug reports for any errors/omissions in the specification on
>>> bugzilla.
>>
>> That's the whole point. It's not clear what the actual specification IS. Is it DMD, dlang.org or TDPL?
>
> The website is the spec.  DMD is the reference instantiation of that spec.  TDPL is a re-framing of it.
>
> Now, given that none of the three are perfect, discrepancies exist which obviously lead to questions.  That's where
> bugzilla comes in to record and track the queue of issues to resolve among them.  Where there's issues, until it's been
> examined, any of the parties could be at fault.
>
> My 2 cents,
> Brad

In all honesty, the website can barely be called a spec. Sure, it has 
grammar and a few notes on semantics, but it's not even close to what 
people expect from a programming language specification in this day and 
age (see C, C++, C#, F#, etc specifications). I'd call it a guide at best.

If someone (probably among the DMD developers) was willing to flesh out 
a clean and complete grammar definition for the language (preferably in 
the form of one EBNF file, independent of doc format), I'd personally be 
willing to put in effort to write a spec around it.

- Alex