What's the current state of D?

[Ary Borenszweig]

Tomas Lindquist Olsen escribió:
> On Fri, May 8, 2009 at 8:56 PM, Walter Bright
> <newshound1 at digitalmars.com> wrote:
>> Steve Teale wrote:
>>> I am not criticizing you. I think you are doing a great job under the
>>> pressure of a slew of suggestions. But maybe a line in the sand at
>>> some point?
>> But there *is* a line in the sand - D1.
>>
>>> OK so for those who crave stability there is D1.x, but when all the
>>> focus appears to be on D2, what level of confidence is afforded to D1
>>> users. Can a project Manager cross his heart and say that D1 will
>>> still be alive and well in five years time?
>> This just makes me very frustrated. Every fix going into D2 that also
>> applies to D1 has been folded into D1. There are even some fixes that only
>> apply to D1 that only go into D1. These average about 20 per month. D1 even
>> got the FreeBSD port before D2. What other language gets that kind of
>> support?
>>
>>
> 
> A lot of good points have been brought forward in this thread. To me
> the big problem with D1 is that we *still* don't have a conforming
> (what does that even mean with D) compiler.
> 
> Yes I'm thinking forward reference problems and such. Sure, these are
> hard problems to fix in the compiler, at least at this point. It's
> something that should have been fixed immediately when the first bug
> report came in.
> 
> From my work on LDC I've seen a lot of the DMD frontend source code,
> its glitches, but of course also those "wow, that's neat" moments.
> However, one thing is very certain, it has some horrible problems with
> order-of-things issues, which the spec says should not matter.
> 
> I'm talking about forward references, differing semantic trees
> depending on the order of declarations, imports etc. even the order in
> which modules are passed on the command line can affect the resulting
> semantic trees (yes I mean AST, but DMD has no distinction, it
> rewrites the AST happily until it thinks the result is sane).
> 
> The main feeling I've come to settle on during the soon two years I've
> spent with LDC, is: we need a new frontend! D1 and D2 both have these
> problems, and they're not going away by themselves.
> 
> -Tomas

That's my feeling also. A new front-end must be done, and the 
specification of the language must be clear in every aspect. No surprises.