{{ message }}
Repository navigation
Debug formatting leads to less desireable output #215
Description
#[macro_use]
extern crate bitflags;
bitflags! {
struct Flags: u32 {
const A = 0b00000001;
const B = 0b00000010;
const C = 0b00000100;
const ABC = Self::A.bits | Self::B.bits | Self::C.bits;
}
}
fn main() {
println!("{:?}", Flags::A | Flags::B | Flags::C );
}prints:
A | B | C | ABCI find it somewhat less helpful that both the expanded (A | B | C) and "compressed" form (ABC) are reported...
Is there a reason behind this? Is this considered more correct for some reason?
Activity
KodrAus commentedon Oct 25, 2021
I think the current algorithm used for debug output is to loop over all flags and append identifiers that correspond to set bits. I think an alternative here that could work would be to short-circuit when we’ve built a format that covers all the set bits. As an implementation note we wouldn’t be able to work off a single source and just mask out bits as we see them, we’ll need to use the whole set of bits to see if a flag is applicable and then a second set that’s masked to see when we’ve catered for all set bits. Otherwise something like this wouldn’t work:
we’d mask out bit 3 for
Abut then have one leftover forBthat doesn’t correspond to it. In that example we’d end up writingA | Bstill, but wouldn’t duplicate compound identifiers the same.In general, I’m not sure if there’s a reasonable algorithm that would produce the smallest possible format for any given set of bits.