On Thursday, 11 June 2015 at 16:33:04 UTC, Dennis Ritchie wrote:
It is possible that Walter and Andrei against macro because of
this:
macro_rules! o_O {
(
$(
$x:expr; [ $( $y:expr ),* ]
);*
) => {
&[ $($( $x + $y ),*),* ]
}
}
fn main() {
let a: &[i32]
= o_O!(10; [1, 2, 3];
20; [4, 5, 6]);
assert_eq!(a, [11, 12, 13, 24, 25, 26]);
}
It looks disgusting! ;)
This baffles me. It seems that language designer always need to
fuck up macros either by:
- Creating a new API to spawn AST, which become a burden on the
compiler development (bonus point if you expose compiler
internal).
- Creating a new syntax, preferably completely inscrutable so
you can pretend you are a guru while using it.
There is a good way to express AST in a language, and this is how
you do it for everything else in the program : you use the damn
language syntax and grammar.
On that one, LISP get it right, except that its general lack of
grammar and syntax (really, it is not that LISP has a lot of (),
it is that everything else has been removed) end up creating a
new set of problems.
It is just this const expr vs CTFE, one has now to learn a new
language to do computation at compile time (const expr) when
using the same language (CTFE) reduce the burden on the dev and
compiler writer.
