how do you define what is "essential stuff", without human input by the author of the crate? what happens when they don't provide that definition? the purpose of a given page of docs is to show everything you can do with a thing, even if that turns out to be a lot
-
-
like, i get it, there's a lot going on, and we've taken strides to streamline this by auto-folding things like trait impls and type definitions but like, my personal opinion is that rustdoc should show you everything that can be known about a type
1 reply 0 retweets 4 likes -
There are a few that I think should be excluded here across the board (as in purely special cased, no hooks) In particular the impls that apply to *every* type (from, into, any at least)
2 replies 0 retweets 1 like -
Replying to @sgrif @QuietMisdreavus and
Can you say more as to why? Users (especially new ones) still need to know they're available
1 reply 0 retweets 2 likes -
Replying to @jntrnr @QuietMisdreavus and
They apply to all types. We should have a "here are the impls that apply for all types" Repeating them on each type is more noise than it's worth imo
1 reply 0 retweets 0 likes -
Replying to @sgrif @QuietMisdreavus and
Thinking about how that would work in practice. User goes to i32 or vector, and looks at the methods it provides. Are you saying don't show the common methods in the list but rather on a separate page?
1 reply 0 retweets 0 likes -
Replying to @jntrnr @QuietMisdreavus and
I'm saying that showing that you can do `let x: i32 = 1i32.into()` is not useful
2 replies 0 retweets 0 likes -
Replying to @sgrif @QuietMisdreavus and
I was asking about where the knowledge of being able to do .into() on the given type should live in the docs. Sounds like you were commenting on something else?
1 reply 0 retweets 0 likes -
Replying to @jntrnr @QuietMisdreavus and
I think showing `Into` impls is super useful, but we don't need to show that a type can be converted into itself over and over again.
2 replies 0 retweets 0 likes -
in a generic context, it’s useful to know that if a function is asking for `T: Into<Thing>`, you can hand it a Thing rustdoc has no concept right now of specialized API pages, so if we filter out these impls, the only place they will show up is on the traits themselves
1 reply 0 retweets 1 like
Right, I get that. I'm not proposing that we generalize this at all. I'm saying that there are a few very specific impls that I think should be special cased and excluded. I don't think this problem really applies outside of std
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.