www.digitalmars.com         C & C++   DMDScript  

digitalmars.dip.ideas - static switch statement

reply IchorDev <zxinsworld gmail.com> writes:
The equivalent of `static if`, but for `switch` statements.

Is there any reason why we don't already have this?
Jul 27
next sibling parent monkyyy <crazymonkyyy gmail.com> writes:
On Monday, 28 July 2025 at 06:49:17 UTC, IchorDev wrote:

 The equivalent of `static if`, but for `switch` statements.

 Is there any reason why we don't already have this?


Id suggest reusing template specailization logic


```d
template foo(T...){
   static switch(T){
     case(int I,string S){
        foo()=>iota(0,I).each!(()=>s.writeln);
     }
     case(S:int){
        foo(int i)=>i.writeln;
     }
     case(S) if(isNumberic!S){
        foo(S i)=>i.writeln("not int");
}}

unittest{
   foo!(10,"foo");
   foo!int(10);
   foo(13.37);//inferred somehow
}
```
Jul 28
prev sibling parent reply Atila Neves <atila.neves gmail.com> writes:
On Monday, 28 July 2025 at 06:49:17 UTC, IchorDev wrote:

 The equivalent of `static if`, but for `switch` statements.

 Is there any reason why we don't already have this?


Probably because it would require implementing and isn't as 
useful as static if and static foreach.
Jul 31
parent reply IchorDev <zxinsworld gmail.com> writes:
On Thursday, 31 July 2025 at 08:36:56 UTC, Atila Neves wrote:

 On Monday, 28 July 2025 at 06:49:17 UTC, IchorDev wrote:

 Is there any reason why we don't already have this?


 Probably because it would require implementing [...]


Could it re-use the logic of a static if-else chain?
Jul 31
parent reply Nick Treleaven <nick geany.org> writes:
On Thursday, 31 July 2025 at 10:12:35 UTC, IchorDev wrote:

 On Thursday, 31 July 2025 at 08:36:56 UTC, Atila Neves wrote:

 On Monday, 28 July 2025 at 06:49:17 UTC, IchorDev wrote:

 Is there any reason why we don't already have this?


 Probably because it would require implementing [...]


 Could it re-use the logic of a static if-else chain?


It could, but then you couldn't do what monkyyy wants, and it 
wouldn't be particularly different from just using a static 
if-else chain except it would be less useful because it couldn't 
use `is` expressions. There's also the smaller issue of whether 
the syntax uses `break;` (which might look odd) or `static break`.

If we get a `match` statement for pattern matching, then it would 
be more useful to have `static match` than `static switch`. But 
also probably harder to implement ;-)
Aug 01
next sibling parent monkyyy <crazymonkyyy gmail.com> writes:
On Friday, 1 August 2025 at 10:32:19 UTC, Nick Treleaven wrote:

 you couldn't do what monkyyy wants


Id rather have bool pattern matching runtime and ct; I suggest it 
because I expect to will lower well

```d
static switch(T){
   case(int I:1){...}
   case(int I:2){...}
   case(int I:3){...}
}
```

which is verbose, repeating the int I over and over again.

but I expect it will lower to:

```d
mixin template __anonymous1(int I:1){...}
mixin template __anonymous2(int I:2){...}
mixin template __anonymous3(int I:3){...}

static if(__traits(compiles,__anonymous1!(T))){ mixin 
__anonymous1!T;} else {
static if(__traits(compiles,__anonymous2!(T))){ mixin 
__anonymous2!T;} else {
static if(__traits(compiles,__anonymous3!(T))){ mixin 
__anonymous3!T;} else {
static assert(false, T.stringof~"failed to match");
}}}
```
Aug 01
prev sibling parent reply user1234 <user1234 12.de> writes:
On Friday, 1 August 2025 at 10:32:19 UTC, Nick Treleaven wrote:

 There's also the smaller issue of whether the syntax uses 
 `break;` (which might look odd) or `static break`.


Yeah, plus `static goto case` and the problem of `final`, as D 
has a runtime error related to switch cases.
Aug 02
parent IchorDev <zxinsworld gmail.com> writes:
On Sunday, 3 August 2025 at 01:18:58 UTC, user1234 wrote:

 On Friday, 1 August 2025 at 10:32:19 UTC, Nick Treleaven wrote:

 There's also the smaller issue of whether the syntax uses 
 `break;` (which might look odd) or `static break`.


 Yeah, plus `static goto case`


I'd prefer a `static` prefix, rather than have a similar issue to 
`break`s always needing a label inside a `static foreach`.


 and the problem of `final`


`final` could always work the same way as at runtime. `default: 
static assert(0);`


 D has a runtime error related to switch cases.


Yep, just make that a compile error.
Aug 19