formatting with alignment fails about time HOT 5 CLOSED

To be exact, it is a common behavior that formatting implementations discard the width and alignment information:

fn main() {
    println!("{:>11?}", "a\nb\nc"); // not aligned
    println!("{:>11}", format!("{:?}", "a\nb\nc")); // aligned
}

I would argue that this behavior is, while frustrating, justifiable because alignment requires the length of printed string beforehand and this might not be acceptable for many types. (For the case of strftime, I think there is no efficient and seemingly-correct solution; Formatter::pad obviously requires the string length.) But it is frustrating, and I agree that this has to be discussed and at least documented.

from time.

If a type does not support alignment, then println! should fail at compile time IMHO.
Alas, this would not be backwards compatible. So we can only make it give a warning.

from time.

Yes @lifthrasiir is right in that most types do not support the width/alignment of the underlying formatting system, and the best solution to this (as you've found) is to call .to_string() and then pad that accordingly. This is currently as-intended, however, and the documentation should probably happen in std::fmt instead of just in this library, so I'm going to close this issue.

from time.

This is currently as-intended, however

Poop. Ok, that was really frustrating. Maybe a warning? A warning would have saved a lot of angst.

the documentation should probably happen in std::fmt instead of just in this library, so I'm going to close this issue.

Where can I submit a PR for that?

from time.

Maybe a warning? A warning would have saved a lot of angst

Unfortunately there's no way for the formatting infrastructure to know whether a type actually looks at the formatting flags or not (as it's all trait-based).

Where can I submit a PR for that?

A PR editing this file should do the trick!

from time.