www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.learn - Static CT Factory

reply Engine Machine <EM EM.com> writes:
I have a template that is suppose to create a type based on a 
template parameter

static if (T == "int")
{
    auto x = New!int();
}
else static if (T == "double")
{
    auto x = New!double();
}

x = 1.234;


This is just an example, I use custom types.

The static if's prevent x's scope from being after them, even 
though it should work perfectly fine. I can't declare x before 
because I don't know the type.

The point is that I don't want to have to put x = 1.234; in each 
static if block when I should be able to do it perfectly fine 
afterwards.

My types that I'm creating all have a name field and I want to 
set it once after the object is created.

Luckily, all of them inherit from the same type and I can declare 
that before the static if block, but in general, that won't work.

I feel that in this case I feel that the scope of the static if 
should allow things to escape since, well, they are static if's. 
It would be similar to #ifdef.

I realize that D doesn't work this way, I'm asking for a nicer 
solution than having to duplicate the same code(x = 1.234 in this 
example) or having use a base class.
Aug 18 2016
next sibling parent reply Adam D. Ruppe <destructionator gmail.com> writes:
On Friday, 19 August 2016 at 01:10:42 UTC, Engine Machine wrote:

 I feel that in this case I feel that the scope of the static if 
 should allow things to escape since, well, they are static if's.



They do.

What, exactly, did you do and what, exactly did you see as the 
error?
Aug 18 2016
parent Engine Machine <EM EM.com> writes:
On Friday, 19 August 2016 at 01:18:28 UTC, Adam D. Ruppe wrote:

 On Friday, 19 August 2016 at 01:10:42 UTC, Engine Machine wrote:

 I feel that in this case I feel that the scope of the static 
 if should allow things to escape since, well, they are static 
 if's.



 They do.

 What, exactly, did you do and what, exactly did you see as the 
 error?


I get undefined identifier, but when create the variable outside 
it works.

I used Anonymouse code and it works so I guess the issue stems 
from something else.

My code is more complex and the stuff is in a foreach. I don't 
know why it isn't working since the simplified code is nearly 
identical. Maybe a branch wasn't being taken, I thought I covered 
them all.
Aug 18 2016
prev sibling parent reply Anonymouse <asdf asdf.net> writes:
On Friday, 19 August 2016 at 01:10:42 UTC, Engine Machine wrote:

 I have a template that is suppose to create a type based on a 
 template parameter

 static if (T == "int")
 {
    auto x = New!int();
 }
 else static if (T == "double")
 {
    auto x = New!double();
 }

 x = 1.234;


 This is just an example, I use custom types.

 The static if's prevent x's scope from being after them, even 
 though it should work perfectly fine. I can't declare x before 
 because I don't know the type.


I'm not sure I understand. static if shouldn't introduce a new 
scope that way.

https://dpaste.dzfl.pl/7b63a6e52309

Mind that x might be a pointer.
Aug 18 2016
parent reply Engine Machine <EM EM.com> writes:
On Friday, 19 August 2016 at 01:25:10 UTC, Anonymouse wrote:

 On Friday, 19 August 2016 at 01:10:42 UTC, Engine Machine wrote:

 I have a template that is suppose to create a type based on a 
 template parameter

 static if (T == "int")
 {
    auto x = New!int();
 }
 else static if (T == "double")
 {
    auto x = New!double();
 }

 x = 1.234;


 This is just an example, I use custom types.

 The static if's prevent x's scope from being after them, even 
 though it should work perfectly fine. I can't declare x before 
 because I don't know the type.


 I'm not sure I understand. static if shouldn't introduce a new 
 scope that way.

 https://dpaste.dzfl.pl/7b63a6e52309

 Mind that x might be a pointer.


Ok, well, I guess the error comes from something else.
Aug 18 2016
parent reply Basile B. <b2.temp gmx.com> writes:
On Friday, 19 August 2016 at 01:53:22 UTC, Engine Machine wrote:

 On Friday, 19 August 2016 at 01:25:10 UTC, Anonymouse wrote:

 On Friday, 19 August 2016 at 01:10:42 UTC, Engine Machine

 x = 1.234;




 Ok, well, I guess the error comes from something else.


*x = 1.234 for when T verifies is(T == int*) produces an error.

You can put an aditional argument in the function header:

void foo(T)(auto ref T t = T.init)
{
	static if (is(T == int))
	    auto x = new Thing!int;
	else static if (is(T == double))
	    auto x = new Thing!double;
	else
             static assert(false, "incorrect type");
	*x = t;
}
Aug 18 2016
parent Engine Machine <EM EM.com> writes:
On Friday, 19 August 2016 at 02:54:48 UTC, Basile B. wrote:

 On Friday, 19 August 2016 at 01:53:22 UTC, Engine Machine wrote:

 On Friday, 19 August 2016 at 01:25:10 UTC, Anonymouse wrote:

 On Friday, 19 August 2016 at 01:10:42 UTC, Engine Machine

 x = 1.234;




 Ok, well, I guess the error comes from something else.


 *x = 1.234 for when T verifies is(T == int*) produces an error.

 You can put an aditional argument in the function header:

 void foo(T)(auto ref T t = T.init)
 {
 	static if (is(T == int))
 	    auto x = new Thing!int;
 	else static if (is(T == double))
 	    auto x = new Thing!double;
 	else
             static assert(false, "incorrect type");
 	*x = t;
 }


It was because I had a added a new type that I didn't account for 
in the static if chain ;/ The static assert is sort of required. 
We sort of need a static switch with required default so 
forgetting this stuff doesn't break later on(in strange ways) 
when new types are added.
Aug 18 2016