Frits van Bommel Wrote:

> lurk wrote:
> > Jarrett Billingsley Wrote:
> > 
> >> "bearophile" <bearophileHUGS@lycos.com> wrote in message news:fd8gnt$149q$1@digitalmars.com...
> >>> I'd like to see some useful small things be backported from 2.x to 1.x, like foreach(i; 2 .. 8).
> >>>
> >>> (In the Python language the future Python V.2.6 will probably be a release to backport as much as possible).
> >> No, see, the entire purpose of the split between 1.0 and 2.0 is that 1.0 won't get any new features.  It's a done deal.  All new features now go into 2.0.
> >>
> >>
> > and the worst is, that the libraries such as tango etc. offer version 2.0
> 
> ??
> AFAIK Tango only supports 1.x ...

meant to be sarcastic.
2.x is need, but nobody supports it.

lurk wrote:
> Frits van Bommel Wrote:
> 
>> lurk wrote:
>>> and the worst is, that the libraries such as tango etc. offer version 2.0
>> ??
>> AFAIK Tango only supports 1.x ...
> 
> meant to be sarcastic.
> 2.x is need, but nobody supports it.
> 

"Hey guys, let's update Tango to support 2.0; it'll take a while, but it'll be worth it!"

"Yeah!"

[A few months later]

"Ok, we're finally done!  We had to refactor a sh*tload of code, and split a heap of it off into different modules and *damn* do I hate string mixins now, but we're finally done!"

Walter: "const sucks; let's re-design it from scratch."

"Dear God, shoot me now!"

*That's* why libraries don't support D 2.0.  Because it would be silly to support a target that's not only moving, but constantly changing shape. :)

	-- Daniel

Daniel Keep wrote:
> 
> lurk wrote:
>> Frits van Bommel Wrote:
>>
>>> lurk wrote:
>>>> and the worst is, that the libraries such as tango etc. offer version 2.0
>>> ??
>>> AFAIK Tango only supports 1.x ...
>> meant to be sarcastic.
>> 2.x is need, but nobody supports it.
>>
> 
> "Hey guys, let's update Tango to support 2.0; it'll take a while, but
> it'll be worth it!"
> 
> "Yeah!"
> 
> [A few months later]
> 
> "Ok, we're finally done!  We had to refactor a sh*tload of code, and
> split a heap of it off into different modules and *damn* do I hate
> string mixins now, but we're finally done!"
> 
> Walter: "const sucks; let's re-design it from scratch."
> 
> "Dear God, shoot me now!"
> 
> *That's* why libraries don't support D 2.0.  Because it would be silly
> to support a target that's not only moving, but constantly changing
> shape. :)
> 
> 	-- Daniel

Which is exactly why cherry picking some backward-compatible features from 2.0 to backport makes a lot of sense.  Is that the point you were trying to make?

--bb

Bill Baxter wrote:
> Daniel Keep wrote:
>>
>> lurk wrote:
>>> Frits van Bommel Wrote:
>>>
>>>> lurk wrote:
>>>>> and the worst is, that the libraries such as tango etc. offer version 2.0
>>>> ??
>>>> AFAIK Tango only supports 1.x ...
>>> meant to be sarcastic.
>>> 2.x is need, but nobody supports it.
>>>
>>
>> [Silliness]
>>
>>     -- Daniel
> 
> Which is exactly why cherry picking some backward-compatible features from 2.0 to backport makes a lot of sense.  Is that the point you were trying to make?
> 
> --bb

No, the point was that "nobody" supports 2.0 because no-one wants to try and support a rapidly moving target.

As for backporting; in general, I would agree.  If DMD was being written by a team of programmers, I'd be pushing for the backporting of every non-compatibility breaking feature that's proven to be useful.

But it's not.  It's just Walter, and Walter can only do so much in a finite amount of time.  I would kill to have some of the 2.0 features in 1.x, but I think having a *stable* compiler is much more important.

I think that's the point *you* were making earlier :P

Incidentally, it's nice to have a non-moving language.  D 1.0 is *very* usable at the moment, and is certainly nicer to program in than C or C++.

	-- Daniel