Document method-call expressions in pseudo-code as well as prose #1432
Conversation
madsmtm
force-pushed
the
method-call-expr-pseudo-code
branch
from
0ddce00 to
4d80b79
Compare
|
|
||
| // Then, for each candidate `T`, add `&T` and `&mut T` to the list | ||
| // immediately after `T`. | ||
| let candidate_receiver_types = candidate_receiver_types.map(|T| [T, &T, &mut T]).flatten(); |
There was a problem hiding this comment.
flat_map? I know this is just pseudo-Rust!
traviscross
added
the
S-waiting-on-review
|
@rustbot labels -S-waiting-on-review +S-wating-on-author We discussed this in the rustdocs call today. On review, we're just not happy with the pseudocode here. It feels like it takes more effort to understand the pseudocode than the value that it's providing. Maybe another way to approach this would be more of an English pseudocode similar to what @ehuss wrote here: It might also be interesting if this pseudocode could be made non-pseudo. In doing that, it would be similar in spirit to:
We're still working out how exactly to tie those things into the Reference/Specification. But this, expanded to be runnnable, would seem to fall in the same category. |
traviscross
added
S-waiting-on-author
|
Yeah, that's fair, I wasn't really happy with it myself, which is why I hadn't pursued this further. Thanks for taking the time to review it, I might still try to fix the linked issues, but will do so in another PR in any case. |
I personally find the documentation on method-call expressions very hard to grok, and I suspect I'm not the only one, so I thought I would attempt to rewrite the section in pseudo-code, to hopefully make the algorithm clearer.
This is done as a precursor to fixing #718 and #1018, which is important for RFC 3519 where we're trying to change the algorithm.