www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.learn - Detecting if a class type (which may or may not have a default

reply "Jarrett Billingsley" <jarrett.billingsley gmail.com> writes:
I just don't think it's possible.  If all classes had default ctors,
it'd be easy; is(typeof(new T)) would be false if and only if T were
abstract.  But since that's not the case, I can't think of a way to
generically see if a given class type is abstract.  Any ideas?

It's always a little frustrating when doing type introspection and
having to rely on weird side-effects and properties of types, when the
compiler is just keeping it in some flag or field somewhere.  Sigh.
"is(T == abstract)"?  :P
Nov 13 2008
next sibling parent BCS <ao pathlink.com> writes:
Reply to Jarrett,


 I just don't think it's possible.  If all classes had default ctors,
 it'd be easy; is(typeof(new T)) would be false if and only if T were
 abstract.  But since that's not the case, I can't think of a way to
 generically see if a given class type is abstract.  Any ideas?
 
 It's always a little frustrating when doing type introspection and
 having to rely on weird side-effects and properties of types, when the
 compiler is just keeping it in some flag or field somewhere.  Sigh.
 "is(T == abstract)"?  :P
 



_traits(isAbstractClass, Class) // 2.0 only IIRC
Nov 13 2008
prev sibling next sibling parent "Jarrett Billingsley" <jarrett.billingsley gmail.com> writes:
On Thu, Nov 13, 2008 at 5:26 PM, BCS <ao pathlink.com> wrote:

 Reply to Jarrett,


 I just don't think it's possible.  If all classes had default ctors,
 it'd be easy; is(typeof(new T)) would be false if and only if T were
 abstract.  But since that's not the case, I can't think of a way to
 generically see if a given class type is abstract.  Any ideas?

 It's always a little frustrating when doing type introspection and
 having to rely on weird side-effects and properties of types, when the
 compiler is just keeping it in some flag or field somewhere.  Sigh.
 "is(T == abstract)"?  :P



 _traits(isAbstractClass, Class) // 2.0 only IIRC



D1.
Nov 13 2008
prev sibling next sibling parent reply Christopher Wright <dhasenan gmail.com> writes:
Jarrett Billingsley wrote:

 I just don't think it's possible.  If all classes had default ctors,
 it'd be easy; is(typeof(new T)) would be false if and only if T were
 abstract.  But since that's not the case, I can't think of a way to
 generically see if a given class type is abstract.  Any ideas?
 
 It's always a little frustrating when doing type introspection and
 having to rely on weird side-effects and properties of types, when the
 compiler is just keeping it in some flag or field somewhere.  Sigh.
 "is(T == abstract)"?  :P



If you know the constructor arguments in advance, you can do something like:
static if (is (typeof (new Foo (1, "hello")))){}

Unfortunately, ParameterTupleOf!(T._ctor) doesn't work:

class AFoo {}
if (is (typeof (AFoo._ctor))) Stdout.formatln ("AFoo._ctor");
if (is (typeof (ParameterTupleOf!(AFoo._ctor)))) Stdout.formatln 
("AFoo._ctor params");
// prints AFoo._ctor

_ctor is a really odd construct -- spotty support, not advertised.
Nov 13 2008
parent Max Samukha <samukha voliacable.com.removethis> writes:
On Thu, 13 Nov 2008 21:07:03 -0500, "Jarrett Billingsley"
<jarrett.billingsley gmail.com> wrote:


On Thu, Nov 13, 2008 at 8:50 PM, Christopher Wright <dhasenan gmail.com> wrote:

 Jarrett Billingsley wrote:

 I just don't think it's possible.  If all classes had default ctors,
 it'd be easy; is(typeof(new T)) would be false if and only if T were
 abstract.  But since that's not the case, I can't think of a way to
 generically see if a given class type is abstract.  Any ideas?

 It's always a little frustrating when doing type introspection and
 having to rely on weird side-effects and properties of types, when the
 compiler is just keeping it in some flag or field somewhere.  Sigh.
 "is(T == abstract)"?  :P



 If you know the constructor arguments in advance, you can do something like:
 static if (is (typeof (new Foo (1, "hello")))){}



Oh, definitely.  But I'm writing a library where the ctor signatures
are provided by the user, and "new T(InitsOf!(Types))" could fail
either because T is abstract or because they just gave an invalid
signature


 Unfortunately, ParameterTupleOf!(T._ctor) doesn't work:

 class AFoo {}
 if (is (typeof (AFoo._ctor))) Stdout.formatln ("AFoo._ctor");
 if (is (typeof (ParameterTupleOf!(AFoo._ctor)))) Stdout.formatln
 ("AFoo._ctor params");
 // prints AFoo._ctor

 _ctor is a really odd construct -- spotty support, not advertised.



I wish it worked right.  Constructors are always the odd ones out.
They're just functions, and should be introspectable as such.



Even D2 doesn't provide a way of getting constructor info except for
the useless "__ctor" returned by __traits(allMembers). It is also
possible to make a useless call to the default constructor in a funky
manner:

class C
{
    this()
    {
        writefln("Ctor");
    }

   this(int x)
   {
   }

    void foo()
    {
        __traits(getMember, C, "__ctor");
    }
}


void main()
{
    auto c = new C;
    c.foo(); // re-constructing the object
}

In other words, there is no introspection for constructors at all.
Nov 14 2008
prev sibling parent "Jarrett Billingsley" <jarrett.billingsley gmail.com> writes:
On Thu, Nov 13, 2008 at 8:50 PM, Christopher Wright <dhasenan gmail.com> wrote:

 Jarrett Billingsley wrote:

 I just don't think it's possible.  If all classes had default ctors,
 it'd be easy; is(typeof(new T)) would be false if and only if T were
 abstract.  But since that's not the case, I can't think of a way to
 generically see if a given class type is abstract.  Any ideas?

 It's always a little frustrating when doing type introspection and
 having to rely on weird side-effects and properties of types, when the
 compiler is just keeping it in some flag or field somewhere.  Sigh.
 "is(T == abstract)"?  :P



 If you know the constructor arguments in advance, you can do something like:
 static if (is (typeof (new Foo (1, "hello")))){}



Oh, definitely.  But I'm writing a library where the ctor signatures
are provided by the user, and "new T(InitsOf!(Types))" could fail
either because T is abstract or because they just gave an invalid
signature


 Unfortunately, ParameterTupleOf!(T._ctor) doesn't work:

 class AFoo {}
 if (is (typeof (AFoo._ctor))) Stdout.formatln ("AFoo._ctor");
 if (is (typeof (ParameterTupleOf!(AFoo._ctor)))) Stdout.formatln
 ("AFoo._ctor params");
 // prints AFoo._ctor

 _ctor is a really odd construct -- spotty support, not advertised.



I wish it worked right.  Constructors are always the odd ones out.
They're just functions, and should be introspectable as such.
Nov 13 2008