• monkyyy (29/29) May 08 2024 Allow saner escape hatches for state-ful templates, instead of
• Atila Neves (3/8) May 10 2024 This is a recipe for bugs wrt differences between compiling
• monkyyy (10/20) May 10 2024 I kinda expect allot of criticism of "you shouldn't even want
• monkyyy (3/6) May 10 2024 or option b) make it per file but id want to make suggestions on

monkyyy <crazymonkyyy gmail.com> writes:

Allow saner escape hatches for state-ful templates, instead of 
compiler ~~bugs~~ fun unintended features with generally horrible 
proforence, syntax and limitations

---

`__COUNTER__` a special token that increases by 1 each time its 
called

```
unittest{
    __COUNTER__.writeln;//0
    __COUNTER__.writeln;//1
}
unittest{
    __COUNTER__.writeln;//2
}
```

this is already possible with mixin and mix files

---

`__traits(oldarguments)` returns the previous template arguments

```
template foo(T...){
   alias bar=__traits(oldarguments);
}
alias a=foo!int;//bar=()
alias b=foo!bool;//bar=(int)
alias c=foo!(float,double);//bar=(int,bool)
alias d=foo!();//bar=(int,bool,float,double)
```

this is already possible with my favorite compiler bug, but with 
n^2 complexity

Atila Neves <atila.neves gmail.com> writes:

On Wednesday, 8 May 2024 at 14:56:37 UTC, monkyyy wrote:
 Allow saner escape hatches for state-ful templates, instead of 
 compiler ~~bugs~~ fun unintended features with generally 
 horrible proforence, syntax and limitations

 ---

 `__COUNTER__` a special token that increases by 1 each time its

This is a recipe for bugs wrt differences between compiling 
separately or not.

monkyyy <crazymonkyyy gmail.com> writes:

On Friday, 10 May 2024 at 13:29:24 UTC, Atila Neves wrote:
 On Wednesday, 8 May 2024 at 14:56:37 UTC, monkyyy wrote:
 Allow saner escape hatches for state-ful templates, instead of 
 compiler ~~bugs~~ fun unintended features with generally 
 horrible proforence, syntax and limitations

 ---

 `__COUNTER__` a special token that increases by 1 each time its

 This is a recipe for bugs wrt differences between compiling 
 separately or not.

I kinda expect allot of criticism of "you shouldn't even want 
this" but my response will be "its already possible, people just 
cant read it"

I believe my `mixin("__LINE__")` method to make a counter will 
also break with separate compilation so _such things are already 
in the compiler_. If this is the main criticism you have, maybe 
it would be possible to add a field to .o files that sets the 
starting point of counter, so if the counter for foo.o ended at 3 
and you compile `dmd bar.d foo.o` `bar.__COUNTER__` is 4.

monkyyy <crazymonkyyy gmail.com> writes:

On Friday, 10 May 2024 at 13:29:24 UTC, Atila Neves wrote:
 
 This is a recipe for bugs wrt differences between compiling 
 separately or not.

or option b) make it per file but id want to make suggestions on 
making edge cases go