Is there a way to find out which class an object belongs to?  If not, how about obj.class?

BTW, I think there is a buglet in the doc:

  D does not have a Java style instanceof operator,
  because the cast operator performs the same function:

          Java:
                  if (a instanceof B)
          D:
                  if ((B) a)

should be:        if (cast(B) a)

1) Yes, there will be a .class or some such. -Walter

2) I am undecided if the cast syntax should be like C or like cast(B) a.

"James Gilbert" <jgrg@sanger.ac.uk> wrote in message news:3B9DF36F.81584272@sanger.ac.uk...
>
>
> Is there a way to find out which class an object belongs to?  If not, how about obj.class?
>
> BTW, I think there is a buglet in the doc:
>
>   D does not have a Java style instanceof operator,
>   because the cast operator performs the same function:
>
>           Java:
>                   if (a instanceof B)
>           D:
>                   if ((B) a)
>
> should be:        if (cast(B) a)

Walter wrote:
> 
> 1) Yes, there will be a .class or some such. -Walter

Great.

> 2) I am undecided if the cast syntax should be like C or like cast(B) a.

I like the explicit "cast".  I thought you were sold
on it from the doc, where you say it removes context-
sensitive parsing of "()".

	James

> "James Gilbert" <jgrg@sanger.ac.uk> wrote in message news:3B9DF36F.81584272@sanger.ac.uk...
> >
> >
> > Is there a way to find out which class an object belongs to?  If not, how about obj.class?
> >
> > BTW, I think there is a buglet in the doc:
> >
> >   D does not have a Java style instanceof operator,
> >   because the cast operator performs the same function:
> >
> >           Java:
> >                   if (a instanceof B)
> >           D:
> >                   if ((B) a)
> >
> > should be:        if (cast(B) a)

In article <3B9DF36F.81584272@sanger.ac.uk>, "James Gilbert" <jgrg@sanger.ac.uk> wrote:

> Is there a way to find out which class an object belongs to?  If not, how about obj.class?
> 
> BTW, I think there is a buglet in the doc:
> 
>   D does not have a Java style instanceof operator, because the cast
>   operator performs the same function:
> 
>           Java:
>                   if (a instanceof B)
>           D:
>                   if ((B) a)
> 
> should be:        if (cast(B) a)

Doesn't this have an ambiguity if B is a boolean type variable (or int if there is no boolean)?  I.e., we could be asking "Is a of type B?" or "Is a true when cast to type B?"


Incidentally, I noted elsewhere that C's -> notation might not be needed
in D.   If you don't mind using it for something else, then casts
would be a possible use:

  foo->long

You could even introduce a new operator, =>, for instanceof so that the following are different:

  if (a->int)...
  if (a=>int)...


(Properties would be another possible use:  myvar->size)

type?  Like Foo'class or Array'length.  This would provide the distinction between class members and primitive type information.  There could be a Foo.class without there being any confusion.  Just thought I'd offer my two cents.

Walter wrote:

> 1) Yes, there will be a .class or some such. -Walter
>
> 2) I am undecided if the cast syntax should be like C or like cast(B) a.
>
> "James Gilbert" <jgrg@sanger.ac.uk> wrote in message news:3B9DF36F.81584272@sanger.ac.uk...
> >
> >
> > Is there a way to find out which class an object belongs to?  If not, how about obj.class?
> >
> > BTW, I think there is a buglet in the doc:
> >
> >   D does not have a Java style instanceof operator,
> >   because the cast operator performs the same function:
> >
> >           Java:
> >                   if (a instanceof B)
> >           D:
> >                   if ((B) a)
> >
> > should be:        if (cast(B) a)

oops, sorry I guess I lost the first line.  Here it is again:

Why don't we use a tick operator ' to retrieve data from a primitive
type?  Like Foo'class or Array'length.  This would provide the distinction
between class members and primitive type information.  There could be a
Foo.class without there being any confusion.  Just thought I'd offer my two
cents.

Erik Rounds wrote:

> type?  Like Foo'class or Array'length.  This would provide the distinction between class members and primitive type information.  There could be a Foo.class without there being any confusion.  Just thought I'd offer my two cents.
>
> Walter wrote:
>
> > 1) Yes, there will be a .class or some such. -Walter
> >
> > 2) I am undecided if the cast syntax should be like C or like cast(B) a.
> >
> > "James Gilbert" <jgrg@sanger.ac.uk> wrote in message news:3B9DF36F.81584272@sanger.ac.uk...
> > >
> > >
> > > Is there a way to find out which class an object belongs to?  If not, how about obj.class?
> > >
> > > BTW, I think there is a buglet in the doc:
> > >
> > >   D does not have a Java style instanceof operator,
> > >   because the cast operator performs the same function:
> > >
> > >           Java:
> > >                   if (a instanceof B)
> > >           D:
> > >                   if ((B) a)
> > >
> > > should be:        if (cast(B) a)

Erik Rounds wrote:
> 
> oops, sorry I guess I lost the first line.  Here it is again:
> 
> Why don't we use a tick operator ' to retrieve data from a primitive
> type?  Like Foo'class or Array'length.  This would provide the distinction
> between class members and primitive type information.  There could be a
> Foo.class without there being any confusion.  Just thought I'd offer my two
> cents.

My $0.02 is that the tick operator plays merry hell
with syntax highlighting in editors.

	James

Ben Cohen wrote:
> 
> In article <3B9DF36F.81584272@sanger.ac.uk>, "James Gilbert" <jgrg@sanger.ac.uk> wrote:
> 
> > Is there a way to find out which class an object belongs to?  If not, how about obj.class?
> >
> > BTW, I think there is a buglet in the doc:
> >
> >   D does not have a Java style instanceof operator, because the cast
> >   operator performs the same function:
> >
> >           Java:
> >                   if (a instanceof B)
> >           D:
> >                   if ((B) a)
> >
> > should be:        if (cast(B) a)
> 
> Doesn't this have an ambiguity if B is a boolean type variable (or int if there is no boolean)?  I.e., we could be asking "Is a of type B?" or "Is a true when cast to type B?"

Good point.

I was also thinking that an "instanceof" operator
might be able to do less work than a cast to get
you the answer.  I suppose the compiler could work
out that the cast was being used in boolean context,
but this would be surpising to the programmer, and
deviate from Walter's idea of a context-free
compiler.

> Incidentally, I noted elsewhere that C's -> notation might not be needed
> in D.   If you don't mind using it for something else, then casts
> would be a possible use:
> 
>   foo->long
> 
> You could even introduce a new operator, =>, for instanceof so that the following are different:
> 
>   if (a->int)...
>   if (a=>int)...
> 
> (Properties would be another possible use:  myvar->size)

I would try to avoid operators that are used for
different things in similar languages!

In article <3BA08FBC.158C9162@sanger.ac.uk>, "James Gilbert" <jgrg@sanger.ac.uk> wrote:

>> Incidentally, I noted elsewhere that C's -> notation might not be needed in D.   If you don't mind using it for something else, then casts would be a possible use:
>> 
>>   foo->long
>> 
...
> I would try to avoid operators that are used for different things in similar languages!

That is sensible ... but a pity to leave the notation unused!

I'm sure that editors can be made to accomidate.  Anyway, the point is having an operator which won't be confused with the . or -> operator.  I know that the -> operator won't be used in D, but I think we should avoid using it in order to avoid confusing C++ programmers new to the language.

James Gilbert wrote:

> Erik Rounds wrote:
> >
> > oops, sorry I guess I lost the first line.  Here it is again:
> >
> > Why don't we use a tick operator ' to retrieve data from a primitive
> > type?  Like Foo'class or Array'length.  This would provide the distinction
> > between class members and primitive type information.  There could be a
> > Foo.class without there being any confusion.  Just thought I'd offer my two
> > cents.
>
> My $0.02 is that the tick operator plays merry hell
> with syntax highlighting in editors.
>
>         James