www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - isRef, isLazy, isOut

reply Eldar Insafutdinov <e.insafutdinov gmail.com> writes:
Following the commits to the dmd repository I found these traits implemented.
The idea is good, but the API chosen does not look right.  Why not have
__traits(storageClass, symbol) and a enum which defines the return value:
enum StorageClass {
    None,
    Ref,
    Lazy,
    Out
}

One could write templates isRef, isLazy etc on top of that if he/she really
needs. IMO this is a cleaner way to implement this particular feature.
Dec 19 2009
parent reply Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
Eldar Insafutdinov wrote:

 Following the commits to the dmd repository I found these traits implemented.
The idea is good, but the API chosen does not look right.  Why not have
__traits(storageClass, symbol) and a enum which defines the return value:
 enum StorageClass {
     None,
     Ref,
     Lazy,
     Out
 }
 
 One could write templates isRef, isLazy etc on top of that if he/she really
needs. IMO this is a cleaner way to implement this particular feature.


I asked in vain for is(var == ref), is(var == out) etc.

Andrei
Dec 19 2009
parent reply Walter Bright <newshound1 digitalmars.com> writes:
Andrei Alexandrescu wrote:

 I asked in vain for is(var == ref), is(var == out) etc.


That's because 'is' works with types, not symbols. It didn't fit.
Dec 19 2009
next sibling parent reply Chris Nicholson-Sauls <ibisbasenji gmail.com> writes:
Walter Bright wrote:

 Andrei Alexandrescu wrote:

 I asked in vain for is(var == ref), is(var == out) etc.


 
 That's because 'is' works with types, not symbols. It didn't fit.


is(typeof(var) == ref) ?

-- Chris Nicholson-Sauls
Dec 20 2009
parent reply Walter Bright <newshound1 digitalmars.com> writes:
Chris Nicholson-Sauls wrote:

 Walter Bright wrote:

 Andrei Alexandrescu wrote:

 I asked in vain for is(var == ref), is(var == out) etc.


 That's because 'is' works with types, not symbols. It didn't fit.


 
 is(typeof(var) == ref) ?


No, because ref is not part of the type. It's a storage class.
Dec 20 2009
parent reply Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
Walter Bright wrote:

 Chris Nicholson-Sauls wrote:

 Walter Bright wrote:

 Andrei Alexandrescu wrote:

 I asked in vain for is(var == ref), is(var == out) etc.


 That's because 'is' works with types, not symbols. It didn't fit.


 is(typeof(var) == ref) ?


 
 No, because ref is not part of the type. It's a storage class.


Speaking of isLazy - any fresh ideas on dissing lazy? FWIW we can 
replace it with the conversion of expressions to delegates. The 
resulting situation is liable to a few ambiguities but is much better 
than lazy.

Andrei
Dec 20 2009
parent Kasumi Hanazuki <k.hanazuki gmail.com> writes:
(2009/12/21 4:43), Andrei Alexandrescu wrote:

 Speaking of isLazy - any fresh ideas on dissing lazy? FWIW we can
 replace it with the conversion of expressions to delegates. The
 resulting situation is liable to a few ambiguities but is much better
 than lazy.


Nobody seemed to mention in the last thread, but D already has the 
implicit expression-to-delegate conversion in the type-safe variadic
parameters. (only available in the last part of parameters though...)

Can we merge lazy with it?

----

void f(void delegate()[1] gs...) {
   gs[0]();
   gs[0]();
}

void g() {
   writeln(42);
}

void main() {
   f(g());
}
Dec 20 2009
prev sibling parent reply Max Samukha <spambox d-coding.com> writes:
On 19.12.2009 23:00, Walter Bright wrote:

 Andrei Alexandrescu wrote:

 I asked in vain for is(var == ref), is(var == out) etc.


 That's because 'is' works with types, not symbols. It didn't fit.


Will there be a way to get storage classes from function symbols or types?

Currently, we have to use a Signature template that, along with 
parameter and return types, extracts storage classes by parsing the 
value of parameter type tuple's stringof. Ugly and unreliable. We need a 
way to do it without resorting to hacks. Something in this vein:

enum StorageClasses
{
   ...
}

class A
{
  ref int foo(const ref int a, lazy int b) const;
  void bar(out a);
}

assert(__traits(storageClasses, A.foo) == StorageClasses.Const | 
StorageClasses.Ref);
assert(__traits(parameterStorageClasses, A.foo) == [StorageClasses.Const 
| StorageClasses.Ref, StorageClasses.Lazy]);

See also http://d.puremagic.com/issues/show_bug.cgi?id=1818
Dec 20 2009
parent Eldar Insafutdinov <e.insafutdinov gmail.com> writes:
Max Samukha Wrote:


 On 19.12.2009 23:00, Walter Bright wrote:

 Andrei Alexandrescu wrote:

 I asked in vain for is(var == ref), is(var == out) etc.


 That's because 'is' works with types, not symbols. It didn't fit.


 
 Will there be a way to get storage classes from function symbols or types?
 
 Currently, we have to use a Signature template that, along with 
 parameter and return types, extracts storage classes by parsing the 
 value of parameter type tuple's stringof. Ugly and unreliable. We need a 
 way to do it without resorting to hacks. Something in this vein:
 
 enum StorageClasses
 {
    ...
 }
 
 class A
 {
   ref int foo(const ref int a, lazy int b) const;
   void bar(out a);
 }
 
 assert(__traits(storageClasses, A.foo) == StorageClasses.Const | 
 StorageClasses.Ref);
 assert(__traits(parameterStorageClasses, A.foo) == [StorageClasses.Const 
 | StorageClasses.Ref, StorageClasses.Lazy]);
 
 See also http://d.puremagic.com/issues/show_bug.cgi?id=1818
 
 


Please, please...
Dec 20 2009