Ralf Schneider:
> Have a look at Ultimate++: http://www.ultimatepp.org/src$Sql$SqlExp$en-us.html

An interesting toolkit.
If you look at this page:
http://www.ultimatepp.org/www$uppweb$overview$en-us.html
The part titled "Value and Null" shows you that they are re-inventing dynamic typing (== always using a variant-like type) to do GUI programming.

There's even a comparison with D:
http://www.ultimatepp.org/www$uppweb$vsd$en-us.html
With the comment: "Means C++ is still well ahead of D (by 70%) if not being hold back by standard library design and average implementation..."
:-)

Bye,
bearophile

bearophile wrote:
> Ralf Schneider:
>> Have a look at Ultimate++: http://www.ultimatepp.org/src$Sql$SqlExp$en-us.html
> 
> An interesting toolkit.
> If you look at this page:
> http://www.ultimatepp.org/www$uppweb$overview$en-us.html
> The part titled "Value and Null" shows you that they are re-inventing dynamic typing (== always using a variant-like type) to do GUI programming.
> 
> There's even a comparison with D:
> http://www.ultimatepp.org/www$uppweb$vsd$en-us.html
> With the comment: "Means C++ is still well ahead of D (by 70%) if not being hold back by standard library design and average implementation..."
> :-)
> 
> Bye,
> bearophile

U++ is an excellent toolkit, it takes a while to get used to their whole transfer-semantics thing, but once you do it's a pleasure. The code is really short and easy to read.

For GUI app development I'd pick U++/C++ over D any day.

(a few years ago, I implemented a 3D height map editor with support for "infinitely" big terrains, WYSIWYG editing/texturing and a lot of other stuff in U++)

Tomas Lindquist Olsen:
> The code is really short and easy to read. For GUI app development I'd pick U++/C++ over D any day.

Is it possible to design a similar API in D too?

I'll take a better look at U++ then, it seems it's partially free too. Its hash maps seem quite faster than D ones, so I think such design ideas may be used to improve D AAs.

I like the API of the now dead WAX GUI toolkit (that works with Wx): http://zephyrfalcon.org/labs/wax.html http://zephyrfalcon.org/waxapi/index_h.html

Bye,
bearophile

bearophile wrote:
> Tomas Lindquist Olsen:
>> The code is really short and easy to read. For GUI app development I'd pick U++/C++ over D any day.
> 
> Is it possible to design a similar API in D too?
> 
> I'll take a better look at U++ then, it seems it's partially free too. Its hash maps seem quite faster than D ones, so I think such design ideas may be used to improve D AAs.
> 
> I like the API of the now dead WAX GUI toolkit (that works with Wx):
> http://zephyrfalcon.org/labs/wax.html
> http://zephyrfalcon.org/waxapi/index_h.html
> 
> Bye,
> bearophile

First, go get the real benefit from U++, you need to use their IDE. To me this was a real pain in the beginning. But _only for a few weeks_. After that it's really a pleasure.

* It has the best highlighting I've seen in any C++ editor (it highlight's scopes with slightly different background colors).
* The forms editor is a must for GUI apps.
* Internationalisation is trivial with built in features (you have to wrap all strings in _t("hello") etc though)
* nice component based project management.

I could go on and on. This thing just makes gui coding so nice, I can't understand why more people aren't using it...

It kinda funny how the way they do all their magic is by completely ignoring "normal" C++ coding practices, with 'const' and 'mutable' [1] being the most important keywords...

[1]: http://www.ultimatepp.org/srcdoc$Core$pick_$en-us.html - search for '#define pick_ const'

Tomas Lindquist Olsen wrote:
> bearophile wrote:
>> Ralf Schneider:
>>> Have a look at Ultimate++: http://www.ultimatepp.org/src$Sql$SqlExp$en-us.html
>>
>> An interesting toolkit.
>> If you look at this page:
>> http://www.ultimatepp.org/www$uppweb$overview$en-us.html
>> The part titled "Value and Null" shows you that they are re-inventing dynamic typing (== always using a variant-like type) to do GUI programming.
>>
>> There's even a comparison with D:
>> http://www.ultimatepp.org/www$uppweb$vsd$en-us.html
>> With the comment: "Means C++ is still well ahead of D (by 70%) if not being hold back by standard library design and average implementation..."
>> :-)
>>
>> Bye,
>> bearophile
> 
> U++ is an excellent toolkit, it takes a while to get used to their whole transfer-semantics thing, but once you do it's a pleasure. The code is really short and easy to read.
> 
> For GUI app development I'd pick U++/C++ over D any day.
> 
> (a few years ago, I implemented a 3D height map editor with support for "infinitely" big terrains, WYSIWYG editing/texturing and a lot of other stuff in U++)
Doesn't seem to have any garbage collection, though.  Probably a part of where their speed advantage comes from.

guslay wrote:
> For those keeping track...
> 
> http://herbsutter.spaces.live.com/
> 
> Crazy syntax, at least it looks better than boost.
> 


What happens to the outer variables? Is a copy of them created for each evaluated closure object? Or does there exist only a single instance of each, as in D?
And how is the closure object destroyed, since there is no GC? (does one have to assign it to a variable, to later delete it?)

I tried to read the proposal to find an answer, but I couldn't quickly grasp it.


-- 
Bruno Medeiros - MSc in CS/E student
http://www.prowiki.org/wiki4d/wiki.cgi?BrunoMedeiros#D

== Quote from Bruno Medeiros (brunodomedeiros+spam@com.gmail)'s article
> guslay wrote:
> > For those keeping track...
> >
> > http://herbsutter.spaces.live.com/
> >
> > Crazy syntax, at least it looks better than boost.
> >
> What happens to the outer variables? Is a copy of them created for each
> evaluated closure object? Or does there exist only a single instance of
> each, as in D?
> And how is the closure object destroyed, since there is no GC? (does one
> have to assign it to a variable, to later delete it?)
> I tried to read the proposal to find an answer, but I couldn't quickly
> grasp it.

From memory, it looked like variables could either be copied or manipulated by reference, and that the closure object was basically just an opaque struct much like a plain old functor.


Sean

Sean Kelly wrote:
> == Quote from Bruno Medeiros (brunodomedeiros+spam@com.gmail)'s article
>> guslay wrote:
>>> For those keeping track...
>>>
>>> http://herbsutter.spaces.live.com/
>>>
>>> Crazy syntax, at least it looks better than boost.
>>>
>> What happens to the outer variables? Is a copy of them created for each
>> evaluated closure object? Or does there exist only a single instance of
>> each, as in D?
>> And how is the closure object destroyed, since there is no GC? (does one
>> have to assign it to a variable, to later delete it?)
>> I tried to read the proposal to find an answer, but I couldn't quickly
>> grasp it.
> 
> From memory, it looked like variables could either be copied or manipulated
> by reference, and that the closure object was basically just an opaque struct
> much like a plain old functor.
> 
> 
> Sean

Yes, that seems to make sense all around.
I was originally surprised by this announcement because it was not possible for C++ to implement "full" closures without having a GC. So it was either closures with outer variables that became invalid after the enclosing function exited (the D situation until some releases ago), or making copies of outer variables.
Still, altough this closure semantics is not as 100% "full" as other languages (latest D, C#, Lisps, etc.) it's still useful enough for C++, and I'm surprised how relatively cleanly it fit in, unlike several of the other extensions which seem like ugly and crufty additions.

-- 
Bruno Medeiros - MSc in CS/E student
http://www.prowiki.org/wiki4d/wiki.cgi?BrunoMedeiros#D