www.digitalmars.com         C & C++   DMDScript  

D - structs should have more of the power of classes

reply imr1984 <imr1984_member pathlink.com> writes:
ok i see no reason why the with() statement shouldnt work with structs.
Logically it makes more sense for with() to only work with structs because you
tend to access structs using their member variables directly, much more than you
do with classes. 

I also think that structs should be able to inherit from member only base
structs.

And access protection should work for structs too.

So the only advantage that classes should have is polymorphism.

thats what i think anyway
Mar 26 2004
next sibling parent J C Calvarese <jcc7 cox.net> writes:
imr1984 wrote:

 ok i see no reason why the with() statement shouldnt work with structs.
 Logically it makes more sense for with() to only work with structs because you
 tend to access structs using their member variables directly, much more than
you
 do with classes. 


In the future with() might work with structs. I think Walter has 
mentioned that it's possible.


 
 I also think that structs should be able to inherit from member only base
 structs.
 
 And access protection should work for structs too.
 
 So the only advantage that classes should have is polymorphism.
 
 thats what i think anyway
 
 



-- 
Justin
http://jcc_7.tripod.com/d/
Mar 26 2004
prev sibling next sibling parent Stewart Gordon <smjg_1998 yahoo.com> writes:
imr1984 wrote:


 ok i see no reason why the with() statement shouldnt work with structs.
 Logically it makes more sense for with() to only work with structs because you
 tend to access structs using their member variables directly, much more than
you
 do with classes. 


I agree.


 I also think that structs should be able to inherit from member only base
 structs.


<snip>

Agree to the extent that it would save defining unions of structs to 
wrap API structs.

Further, such a struct, in which only methods are added, could be 
implicitly convertible to/from its base.  Whereas one that adds fields 
to the struct (if we should allow such a thing) wouldn't.

Stewart.

-- 
My e-mail is valid but not my primary mailbox, aside from its being the 
unfortunate victim of intensive mail-bombing at the moment.  Please keep 
replies on the 'group where everyone may benefit.
Mar 26 2004
prev sibling parent reply J Anderson <REMOVEanderson badmama.com.au> writes:
imr1984 wrote:


ok i see no reason why the with() statement shouldnt work with structs.
Logically it makes more sense for with() to only work with structs because you
tend to access structs using their member variables directly, much more than you
do with classes. 

I also think that structs should be able to inherit from member only base
structs.

And access protection should work for structs too.

So the only advantage that classes should have is polymorphism.

thats what i think anyway

  


I agree to.  I like structs because they are light.  I hate structs 
because they are not reusable.  I've come up against this problem over 
and over again.

-- 
-Anderson: http://badmama.com.au/~anderson/
Mar 26 2004
parent reply Dave Sieber <dsieber spamnot.sbcglobal.net> writes:
J Anderson <REMOVEanderson badmama.com.au> wrote:


 I agree to.  I like structs because they are light.  I hate structs 
 because they are not reusable.  I've come up against this problem over
 and over again.


In my first programming efforts with D (converting a simple text game), the 
main problem with structs that I ran into was not having constructors.  In 
C++, I routinely used these to initialize structs:

Point p = Point(10, 20);

As I understand D so far (and I am only a grasshopper), this now takes 
three lines:

Point p;
p.x = 10;
p.y = 20;

-- 
dave
Mar 26 2004
next sibling parent J Anderson <REMOVEanderson badmama.com.au> writes:
Dave Sieber wrote:


J Anderson <REMOVEanderson badmama.com.au> wrote:

  


I agree to.  I like structs because they are light.  I hate structs 
because they are not reusable.  I've come up against this problem over
and over again.
    


In my first programming efforts with D (converting a simple text game), the 
main problem with structs that I ran into was not having constructors.  In 
C++, I routinely used these to initialize structs:

Point p = Point(10, 20);

As I understand D so far (and I am only a grasshopper), this now takes 
three lines:

Point p;
p.x = 10;
p.y = 20;
  


This type of construction can be written in a method or a free 
function.  I don't have much of a problem with constructors.  Mainly 
inheritance.

-- 
-Anderson: http://badmama.com.au/~anderson/
Mar 26 2004
prev sibling parent reply "C. Sauls" <ibisbasenji yahoo.com> writes:
Point p = { x:10, y:20 };

-C. Sauls
-Invironz

Dave Sieber wrote:

 J Anderson <REMOVEanderson badmama.com.au> wrote:
 
 

I agree to.  I like structs because they are light.  I hate structs 
because they are not reusable.  I've come up against this problem over
and over again.


 
 
 In my first programming efforts with D (converting a simple text game), the 
 main problem with structs that I ran into was not having constructors.  In 
 C++, I routinely used these to initialize structs:
 
 Point p = Point(10, 20);
 
 As I understand D so far (and I am only a grasshopper), this now takes 
 three lines:
 
 Point p;
 p.x = 10;
 p.y = 20;
Mar 26 2004
parent reply Dave Sieber <dsieber spamnot.sbcglobal.net> writes:
"C. Sauls" <ibisbasenji yahoo.com> wrote:


 Point p = { x:10, y:20 };


Wow! How did I miss that? :-)  Ok, it's only been two days, I have a lot to 
learn.

After I wrote that earlier reply, I realized that it was mostly in function 
calls that I was doing something like that, e.g.

object.moveTo(Point(10, 20));

I was converting some C++ code which used this kind of call frequently. 
Anyway, there are numerous ways to achieve the same result, including 
modifying, in this case, moveTo() to take the arguments directly, as an 
overloaded variant:

object.MoveTo(10, 20);

-- 
dave
Mar 26 2004
parent reply "C. Sauls" <ibisbasenji yahoo.com> writes:
Comments imbedded: (who started this phrase-trend, btw?)

Dave Sieber wrote:

 "C. Sauls" <ibisbasenji yahoo.com> wrote:

Point p = { x:10, y:20 };


 Wow! How did I miss that? :-)  Ok, it's only been two days, I have a lot to 
 learn.


Happens to me constantly, don't worry about it.


 object.moveTo(Point(10, 20));


In that case, you could always do:

<snip>
// ...
struct Point {
   int  x,
        y;

   static Point opCall(int xx, int yy) {
     Point p = { x:xx, y:yy };
     return p;
   }
}
// ...
object.moveTo(Point(10, 20));
// ...
</snip>

I forget who posted this idea before... credit to whomever it was. 
Guess I could be non-lazy and search for it... but alas, lazy I am.

-C. Sauls
-Invironz
Mar 26 2004
parent Dave Sieber <dsieber spamnot.sbcglobal.net> writes:
"C. Sauls" <ibisbasenji yahoo.com> wrote:


 In that case, you could always do:
 
 struct Point {
    int  x,
         y;
 
    static Point opCall(int xx, int yy) {
      Point p = { x:xx, y:yy };
      return p;
    }
 }
 // ...
 object.moveTo(Point(10, 20));
 // ...
 </snip>


Thanks again, another new feature I've learned -- I had used opCall, but 
hadn't thought about using it with a struct.

BTW, one minor point: I can't initialize the Point as you wrote it above, 
the compiler complains that p isn't static. No biggie -- a slight 
adjustment, and it's all working, in particular the moveTo(Point(10, 20)) 
expression.

And -- I got the game (my first project, converting an old C++ app to D) 
converted, compiled, and running! Intense, I've been on it all day (one 
benefit of being unemployed :-)  Many, many compile errors, access 
violations, array bounds errors -- you name it. And a lot of help here, so 
thanks to everyone for being so helpful. Still a couple screen boo-boos in 
the game, but they are cosmetic.

I've also got WinDbg working, to locate the Access Violations, and I wrote 
some logging routines as well, to help track where the problems were 
(couldn't use printf() because of the game screen). A good learning 
experience so far, but I do wish for better debugging. A line number for 
access violations/exceptions would be nice, in debug builds when it is 
available. Would have helped a lot -- we have them for array bounds errors 
and switch/case default errors, so I think it should be doable...?

-- 
dave
Mar 26 2004