www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - Please do not defer the scoped closure

reply Frank Benoit <keinfarbton googlemail.com> writes:
I am a bit nervous about the full closure does the silent heap 
allocation. It was said, that there will be a possibility to make it 
work without the allocation by using 'scope'.

See:
http://www.digitalmars.com/pnews/read.php?server=news.digitalmars.com&group=digitalmars.D&artnum=62542

I have lots of code using delegate literal in realtime code. I use it as 
part of flow control. So refactoring (to avoid the delegate litarals) is 
hard.

Please do not defer that much longer. It's like removing an important 
feature of D1 out of D2.
Mar 21 2008
parent reply "Craig Black" <cblack ara.com> writes:
"Frank Benoit" <keinfarbton googlemail.com> wrote in message 
news:fs0rvu$8hm$1 digitalmars.com...
I am a bit nervous about the full closure does the silent heap allocation. 
It was said, that there will be a possibility to make it work without the 
allocation by using 'scope'.

 See:
 http://www.digitalmars.com/pnews/read.php?server=news.digitalmars.com&group=digitalmars.D&artnum=62542

 I have lots of code using delegate literal in realtime code. I use it as 
 part of flow control. So refactoring (to avoid the delegate litarals) is 
 hard.

 Please do not defer that much longer. It's like removing an important 
 feature of D1 out of D2.
Lambdas syntax in C++ '0X will include a provision to specify explicitly whether to instantiate stack variables on the heap or not. -Craig
Mar 21 2008
parent bearophile <bearophileHUGS lycos.com> writes:
Craig Black:
 Lambdas syntax in C++ '0X will include a provision to specify explicitly 
 whether to instantiate stack variables on the heap or not.
I have found it here: http://en.wikipedia.org/wiki/C%2B%2B0x#Lambda_functions_and_expressions The syntax for simple expressions is shorter than D one: <>(int x, int y) (x + y) In D: (int x, int y) { return x + y; } But reading that Wikipedia page the semantics is really too much complex! Those C++0x designers seem mad, they think that the mind of programmers has no bounds. I like languages that leave a bit of free space in my mind, because I have other things too to think about, and other languages to know. For D the default can be the safer one (on the heap), with something optional to use the stack instead. (I haven't used the D 2.x yet, but I'd like to write a tiny monad-based parser with its closures). Bye, bearophile
Mar 21 2008