Remove #[doc(hidden)] from macros in std::prelude.

[m-ou-se]

Seems like this rustdoc bug was fixed.

Fixes #63268.
cc rust-lang/rfcs#3090 (comment)

r? @jyn514

[m-ou-se]

I just noticed that some of the prelude imports are documented as crate:: and some as core::prelude::v1::: https://doc.rust-lang.org/1.50.0/std/prelude/v1/index.html

This just matches how the paths are expressed in code, but it means that the traits that are now imported together with the identically named macros from the core prelude are also documented as core::prelude::v1:: and not crate:: like the other traits.

[m-ou-se]

Maybe we should just do pub use core::prelude::v1::*; in std instead.

[petrochenkov]

Could you mark this PR as closing #63268?

[jyn514]

Maybe we should just do pub use core::prelude::v1::*; in std instead.

That seems fine to me, but I think it's a T-libs decision. r=me on this specific change, I'll leave it up to you if you want to change it to a glob reexport.

[jyn514]

Could you mark this PR as closing #63268?

Done, good catch!

[rust-log-analyzer]

The job x86_64-gnu-llvm-9 failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)

.................................................................................................... 9300/11557
.................................................................................................... 9400/11557
.....................................................................................i......i....... 9500/11557
.................................................................................................... 9600/11557
...............................iiiiiii..iiiiii.i.................................................... 9700/11557
.................................................................................................... 9900/11557
.................................................................................................... 10000/11557
.................................................................................................... 10100/11557
.................................................................................................... 10200/11557
---
Suite("src/test/assembly") not skipped for "bootstrap::test::Assembly" -- not in ["src/tools/tidy"]
Check compiletest suite=assembly mode=assembly (x86_64-unknown-linux-gnu -> x86_64-unknown-linux-gnu)

running 29 tests
iiiiiiiiiiiiiiiiiiiiiiiiiiiii

Suite("src/test/incremental") not skipped for "bootstrap::test::Incremental" -- not in ["src/tools/tidy"]
 finished in 0.067 seconds
Check compiletest suite=incremental mode=incremental (x86_64-unknown-linux-gnu -> x86_64-unknown-linux-gnu)
---
Suite("src/test/debuginfo") not skipped for "bootstrap::test::Debuginfo" -- not in ["src/tools/tidy"]
Check compiletest suite=debuginfo mode=debuginfo (x86_64-unknown-linux-gnu -> x86_64-unknown-linux-gnu)

running 116 tests
iiiiiiiiii.i.i..i..i....ii...ii....ii..........iiii.........i.....i...i.......ii.i.ii.....iiii.....i 100/116
test result: ok. 78 passed; 0 failed; 38 ignored; 0 measured; 0 filtered out; finished in 2.45s

 finished in 2.518 seconds
Suite("src/test/ui-fulldeps") not skipped for "bootstrap::test::UiFullDeps" -- not in ["src/tools/tidy"]
---
Building stage0 tool linkchecker (x86_64-unknown-linux-gnu)
   Compiling once_cell v1.4.1
   Compiling linkchecker v0.1.0 (/checkout/src/tools/linkchecker)
    Finished release [optimized] target(s) in 2.06s
std/prelude/v1/index.html:7: broken link - core/macros/builtin/attr.bench.html
std/prelude/v1/index.html:7: broken link - core/macros/builtin/attr.global_allocator.html
std/prelude/v1/index.html:7: broken link - core/macros/builtin/attr.test.html
std/prelude/v1/index.html:7: broken link - core/macros/builtin/attr.test_case.html
std/prelude/v1/index.html:7: broken link - core/macros/builtin/derive.RustcDecodable.html
std/prelude/v1/index.html:7: broken link - core/macros/builtin/derive.RustcEncodable.html
thread 'main' panicked at 'found some broken links', src/tools/linkchecker/main.rs:97:9


command did not execute successfully: "/checkout/obj/build/x86_64-unknown-linux-gnu/stage0-tools-bin/linkchecker" "/checkout/obj/build/x86_64-unknown-linux-gnu/doc"
expected success, got: exit code: 101

[m-ou-se]

Looks like it doesn't work yet for the few macros that are not available anywhere in std's public interface (other than the prelude).

[jyn514]

Looks like it doesn't work yet for the few macros that are not available anywhere in std's public interface (other than the prelude).

Could you change it so only those macros are marked as doc(hidden) in the meantime?

[jyn514]

Also, if you mark those as doc(inline), does it show the documentation it should?

[m-ou-se]

I just noticed that some of the prelude imports are documented as crate:: and some as core::prelude::v1::: https://doc.rust-lang.org/1.50.0/std/prelude/v1/index.html

This just matches how the paths are expressed in code, but it means that the traits that are now imported together with the identically named macros from the core prelude are also documented as core::prelude::v1:: and not crate:: like the other traits.

Trying out some stuff.. It's quite tricky/impossible to express them all in terms of crate:: in std, because we want the Debug macro, but not the Debug trait, but they're in the same path in std. We also want std::env the macro, but not std::env the module, etc.

[m-ou-se]

Could you change it so only those macros are marked as doc(hidden) in the meantime?

That's exactly the state before this PR.

Looks like the change for the ones that do work already happened in aa863ca (which seems to have the wrong commit message), as part of #77862

[m-ou-se]

Also, if you mark those as doc(inline), does it show the documentation it should?

Yes, that works fine. (It does break the link to std::alloc::GlobalAlloc in the inlined global_allocator's documentation though, as that's just a hardcoded relative url from core to std.)

[jyn514]

Also, if you mark those as doc(inline), does it show the documentation it should?

Yes, that works fine. (It does break the link to std::alloc::GlobalAlloc in the inlined global_allocator's documentation though, as that's just a hardcoded relative url from core to std.)

That should be fixed by using ../../std (instead of ../std).

[m-ou-se]

Closing this, as the issue with PartialOrd, etc. was already solved, and doc(inline)'ing the remaining ones is probably not a great solution. Might open a PR at some point to change the entire std prelude to just glob-import alloc and core's preludes.