www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.learn - Miscelleanous Questions

reply Bruno Medeiros <daiphoenixNO SPAMlycos.com> writes:
Miscelleanous questions, here is a load of them. Sorry if some may not 
be be truly D specific. :p

Is static initialization of fixed-length arrays only permited for 
static-allocation fixed-length arrays? (what a name mess :o ) Why so?
Also is it possible to declare a fixed-length array without directly 
specifying the size, but instead determining it by the size of the 
static initializer? Like the following in C:
int iar[] = { 1, 2, 8, 7, 2 };   // iar is a fixed-length array of size 
5 elements


Is there an add/insert function for associative arrays (like there is 
remove() )? I think I rather prefer it than the "aa[key] = value;" way.


CLASSES: "3 It is illegal to refer to this implicitly or explicitly 
prior to making a constructor call."
Huh, how is that even possible to do?


CLASSES: "There can be only one destructor per class, the destructor 
does not have any parameters, and has no attributes. It is always virtual."
Why is a destructor allways virtual?


MEMORY MANAGEMENT: "The pointer returned from new() must be to memory 
aligned to the default alignment. This is 8 on win32 systems."
8? On win32 is the alignment not 4 for pointers?


MEMORY MANAGEMENT: "The critical features of delete() are: ...
The pointer p may be null."
What? What sense does it make, deletion with a null pointer?



-- 
Bruno Medeiros
Computer Science/Engineering student
Aug 09 2005
next sibling parent reply "Jarrett Billingsley" <kb3ctd2 yahoo.com> writes:
"Bruno Medeiros" <daiphoenixNO SPAMlycos.com> wrote in message 
news:ddb14r$b60$1 digitaldaemon.com...

 Is static initialization of fixed-length arrays only permited for 
 static-allocation fixed-length arrays? (what a name mess :o ) Why so?
 Also is it possible to declare a fixed-length array without directly 
 specifying the size, but instead determining it by the size of the static 
 initializer? Like the following in C:
 int iar[] = { 1, 2, 8, 7, 2 };   // iar is a fixed-length array of size 5 
 elements


Array literals for non-static arrays aren't supported yet, unfortunately. 
And I don't think you can make a fixed-length array like that.  You can do 
this:

static int[] blah=[1,2,3];

But it'll be a dynamic array.


 Is there an add/insert function for associative arrays (like there is 
 remove() )? I think I rather prefer it than the "aa[key] = value;" way.


I agree with you there, and there has been A LOT of debate over AAs.  I 
think an "add" or "put" function would be very welcome.


 CLASSES: "3 It is illegal to refer to this implicitly or explicitly prior 
 to making a constructor call."
 Huh, how is that even possible to do?


Maybe it's not ;)


 CLASSES: "There can be only one destructor per class, the destructor does 
 not have any parameters, and has no attributes. It is always virtual."
 Why is a destructor allways virtual?


Because otherwise, you'd get problems when you destroyed classes that 
inherited from other classes.  Whenever you have a derived class, the dtor 
should be virtual.


 MEMORY MANAGEMENT: "The pointer returned from new() must be to memory 
 aligned to the default alignment. This is 8 on win32 systems."
 8? On win32 is the alignment not 4 for pointers?


I don't know about that one.


 MEMORY MANAGEMENT: "The critical features of delete() are: ...
 The pointer p may be null."
 What? What sense does it make, deletion with a null pointer?


Because it's legal.  "delete null" is a valid statement.  In D (and in C++), 
calling "delete null" does nothing instead of causing an access violation.
Aug 09 2005
parent reply Bruno Medeiros <daiphoenixNO SPAMlycos.com> writes:
Jarrett Billingsley wrote:

 "Bruno Medeiros" <daiphoenixNO SPAMlycos.com> wrote in message 

CLASSES: "There can be only one destructor per class, the destructor does 
not have any parameters, and has no attributes. It is always virtual."
Why is a destructor allways virtual?


 
 
 Because otherwise, you'd get problems when you destroyed classes that 
 inherited from other classes.  Whenever you have a derived class, the dtor 
 should be virtual.
 
 


Yes, I understand in that case, but what then if you dont have a derived
  class?..



MEMORY MANAGEMENT: "The critical features of delete() are: ...
The pointer p may be null."
What? What sense does it make, deletion with a null pointer?


 
 
 Because it's legal.  "delete null" is a valid statement.  In D (and in C++), 
 calling "delete null" does nothing instead of causing an access violation. 
 


I was talking about runtime, not compile-time validity. I think it would 
be obvious that "delete null;" would be a compile-time valid statement. 
Turns out we were both wrong and now I really dont know what to think 
anymore: I tested a "delete null;" and it fails compilation with a "null 
is not an lvalue" error, but a "int* pt = null; delete pt;" is 
compilable, and runs without crashing. :S
What's the deal after all?


-- 
Bruno Medeiros
Computer Science/Engineering student
Aug 11 2005
parent reply "Jarrett Billingsley" <kb3ctd2 yahoo.com> writes:
"Bruno Medeiros" <daiphoenixNO SPAMlycos.com> wrote in message 
news:ddgv49$1ooi$1 digitaldaemon.com...

 Yes, I understand in that case, but what then if you dont have a derived
  class?..


Does it much matter?  The performance overhead of a virtual destructor would 
be rather small, I'd think.  Especially since in larger projects, classes 
which don't have any derived classes are rare.


 I was talking about runtime, not compile-time validity. I think it would 
 be obvious that "delete null;" would be a compile-time valid statement. 
 Turns out we were both wrong and now I really dont know what to think 
 anymore: I tested a "delete null;" and it fails compilation with a "null 
 is not an lvalue" error, but a "int* pt = null; delete pt;" is compilable, 
 and runs without crashing. :S
 What's the deal after all?


delete in D does something that it doesn't in C++ - it also sets the pointer 
that was passed in to null.  So writing "delete null" isn't valid, as it'd 
be like trying to pass a number literal to a function that took an inout 
int.
Aug 11 2005
parent Bruno Medeiros <daiphoenixNO SPAMlycos.com> writes:
Jarrett Billingsley wrote:

 "Bruno Medeiros" <daiphoenixNO SPAMlycos.com> wrote in message 
 news:ddgv49$1ooi$1 digitaldaemon.com...
 

Yes, I understand in that case, but what then if you dont have a derived
 class?..


 
 
 Does it much matter?  The performance overhead of a virtual destructor would 
 be rather small, I'd think.  Especially since in larger projects, classes 
 which don't have any derived classes are rare.
 


Well kind of, even if just as a conceptual question.


I was talking about runtime, not compile-time validity. I think it would 
be obvious that "delete null;" would be a compile-time valid statement. 
Turns out we were both wrong and now I really dont know what to think 
anymore: I tested a "delete null;" and it fails compilation with a "null 
is not an lvalue" error, but a "int* pt = null; delete pt;" is compilable, 
and runs without crashing. :S
What's the deal after all?


 
 
 delete in D does something that it doesn't in C++ - it also sets the pointer 
 that was passed in to null.  So writing "delete null" isn't valid, as it'd 
 be like trying to pass a number literal to a function that took an inout 
 int. 
 


Ahh, I see now, that never had occured to me, but it makes perfect sense 
since otherwise there would be GC-related problems.

-- 
Bruno Medeiros
Computer Science/Engineering student
Aug 14 2005
prev sibling parent "Regan Heath" <regan netwin.co.nz> writes:
On Tue, 09 Aug 2005 19:45:29 +0000, Bruno Medeiros  
<daiphoenixNO SPAMlycos.com> wrote:

 Also is it possible to declare a fixed-length array without directly  
 specifying the size, but instead determining it by the size of the  
 static initializer? Like the following in C:
 int iar[] = { 1, 2, 8, 7, 2 };   // iar is a fixed-length array of size  
 5 elements


I don't think so. But if I may suggest a syntax...

int iar[$] = [ 1, 2, 8, 7, 2 ];

Makes sense to me as $ means length in slice context.

Regan
Aug 09 2005