Re: [Texmacs-dev] Efficiency of C++ switch-statements

[Joris van der Hoeven]

> I have made a very small test program:
>
> #include <stdlib.h>
>
> int main(){
>
>   int test;
>
>   test = rand();
>
>   switch(test){
>   case 1:
>   case 10:
>   case 100:
>   case 1000:
>   case 10000:
>   case 100000:
>   case 1000000: test++; break;
>   case 2:
>   case 20:
>   case 200:
>   case 2000:
>   case 20000:
>   case 200000:
>   case 2000000: test++; break;
>   case 3:
>   case 30:
>   case 300:
>   case 3000:
>   case 30000:
>   case 300000:
>   case 3000000: test++; break;
>   case 4: test++; break;
>   case 40: test++; break;
>   default: break;
>   }
> }

OK, but this is not a very dense table.
Could you please test with something like

switch(test){
case 3: ...;
case 6: ...;
case 9: ...;
...
case 27: ...;
case 30: ...;
default: ...;
}

or similar dense tables with some missing (or permuted) items?
The switches are meant to range over an important part (but not all)
TeXmacs primitives.