Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Rework rustdoc const type #82873

Merged
merged 5 commits into from
Mar 26, 2021
Merged

Conversation

GuillaumeGomez
Copy link
Member

@GuillaumeGomez GuillaumeGomez commented Mar 7, 2021

This PR is mostly about two things:

  1. Not storing some information in the clean::Constant type
  2. Using TyCtxt in the formatting (which we will need in any case as we move forward in any case).

Also: I'm very curious of the perf change in here.

Thanks a lot @danielhenrymantilla for your Captures idea! It allowed me to solve the lifetime issue completely. :)

r? @jyn514

@GuillaumeGomez GuillaumeGomez added the T-rustdoc Relevant to the rustdoc team, which will review and decide on the PR/issue. label Mar 7, 2021
@rust-highfive rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Mar 7, 2021
@GuillaumeGomez
Copy link
Member Author

@bors try @rust-timer queue

@rust-timer
Copy link
Collaborator

Awaiting bors try build completion.

@rustbot label: +S-waiting-on-perf

@rustbot rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Mar 7, 2021
@bors
Copy link
Contributor

bors commented Mar 7, 2021

⌛ Trying commit b9c09e42b696c2dc58090555fd7e67592ff2e8af with merge b9ff6ae773d5f0ce233ef2dca5abb531e05c6809...

@bors
Copy link
Contributor

bors commented Mar 7, 2021

☀️ Try build successful - checks-actions
Build commit: b9ff6ae773d5f0ce233ef2dca5abb531e05c6809 (b9ff6ae773d5f0ce233ef2dca5abb531e05c6809)

@rust-timer
Copy link
Collaborator

Queued b9ff6ae773d5f0ce233ef2dca5abb531e05c6809 with parent 66ec64c, future comparison URL.

Copy link
Member

@jyn514 jyn514 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I haven't read the whole thing yet. Could you describe more the motivation for this change? You said

Not storing some information in the clean::Constant type

which makes sense to me - is the idea just to help performance?

src/librustdoc/clean/mod.rs Outdated Show resolved Hide resolved
src/librustdoc/clean/types.rs Outdated Show resolved Hide resolved
src/librustdoc/html/render/mod.rs Outdated Show resolved Hide resolved
@rust-timer
Copy link
Collaborator

Finished benchmarking try commit (b9ff6ae773d5f0ce233ef2dca5abb531e05c6809): comparison url.

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. Please note that if the perf results are neutral, you should likely undo the rollup=never given below by specifying rollup- to bors.

Importantly, though, if the results of this run are non-neutral do not roll this PR up -- it will mask other regressions or improvements in the roll up.

@bors rollup=never
@rustbot label: +S-waiting-on-review -S-waiting-on-perf

@rustbot rustbot removed the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Mar 7, 2021
@GuillaumeGomez
Copy link
Member Author

The benchmark results seem mostly positive (just wondering what's happening with ctfe-stress-4-doc):

Screenshot from 2021-03-07 22-37-53

@GuillaumeGomez
Copy link
Member Author

I haven't read the whole thing yet. Could you describe more the motivation for this change? You said

Not storing some information in the clean::Constant type

which makes sense to me - is the idea just to help performance?

It's definitely targetting performance here. It's also preparing the field in html/format by passing the TyCtxt type around since we'll very likely make more things into the formatting directly too (and this is where most changes actually come from...).

@GuillaumeGomez
Copy link
Member Author

So let's see if your idea will improve the perf, even a little. ;)

@bors try @rust-timer queue

@rust-timer
Copy link
Collaborator

Awaiting bors try build completion.

@rustbot label: +S-waiting-on-perf

@rustbot rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Mar 7, 2021
@bors
Copy link
Contributor

bors commented Mar 7, 2021

⌛ Trying commit 4357e647282efb170f5ada5d43e09ae23eb62b7f with merge d1670c5e02aa01a73205d1a1ff58e5fda1044425...

@jyn514
Copy link
Member

jyn514 commented Mar 7, 2021

It's also preparing the field in html/format by passing the TyCtxt type around since we'll very likely make more things into the formatting directly too (and this is where most changes actually come from...).

Ok, so the TyCtxt isn't actually needed for these changes? Could you split it into a different PR (or just commit) so it's easier to see which changes are necessary?

@GuillaumeGomez
Copy link
Member Author

Well... Actually it is needed because you need it to print the const. And from that, all the "upper" functions needed it too. You'll note that not all functions in html/format take a TyCtxt, that's the reason. I "only" changed the ones which needed it because one of the children was using it. Pretty impressive how one change is affecting so many things. If it wasn't needed yet, I wouldn't have been able to add a parameter without using underscore to prevent warnings. ;)

@bors
Copy link
Contributor

bors commented Mar 7, 2021

☀️ Try build successful - checks-actions
Build commit: d1670c5e02aa01a73205d1a1ff58e5fda1044425 (d1670c5e02aa01a73205d1a1ff58e5fda1044425)

@rust-timer
Copy link
Collaborator

Queued d1670c5e02aa01a73205d1a1ff58e5fda1044425 with parent 66ec64c, future comparison URL.

@rust-timer
Copy link
Collaborator

Finished benchmarking try commit (d1670c5e02aa01a73205d1a1ff58e5fda1044425): comparison url.

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. Please note that if the perf results are neutral, you should likely undo the rollup=never given below by specifying rollup- to bors.

Importantly, though, if the results of this run are non-neutral do not roll this PR up -- it will mask other regressions or improvements in the roll up.

@bors rollup=never
@rustbot label: +S-waiting-on-review -S-waiting-on-perf

@rustbot rustbot removed the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Mar 8, 2021
@GuillaumeGomez
Copy link
Member Author

No notable changes since before. Waiting for the review then!

Copy link
Member

@jyn514 jyn514 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The benchmark results seem mostly positive (just wondering what's happening with ctfe-stress-4-doc):

Wow, that's amazing! I looked at the self-profile data and it looks like we've almost completely eliminated the time eval_to_allocation_raw takes to execute (!!). Do you know how that happened? I don't see it being called in rustdoc directly.

eval_to_allocation_raw | 0.002 | 0.39% | -99.9% (-1.191) | 151 | -5.0% (-8.000)

I haven't looked at the rest of the changes in detail but this definitely seems worth landing if it has that big an impact :)

src/librustdoc/clean/types.rs Outdated Show resolved Hide resolved
src/librustdoc/clean/types.rs Outdated Show resolved Hide resolved
src/librustdoc/clean/types.rs Outdated Show resolved Hide resolved
src/librustdoc/clean/types.rs Outdated Show resolved Hide resolved
@jyn514 jyn514 added the relnotes-perf Performance improvements that should be mentioned in the release notes. label Mar 8, 2021
@bors
Copy link
Contributor

bors commented Mar 22, 2021

☔ The latest upstream changes (presumably #82855) made this pull request unmergeable. Please resolve the merge conflicts.

@GuillaumeGomez
Copy link
Member Author

r? @ollie27

@rust-highfive rust-highfive assigned ollie27 and unassigned jyn514 Mar 22, 2021
Copy link
Member

@jyn514 jyn514 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think I'm ready to approve this with a couple nits fixed. I don't feel confident I've caught all the bugs, but hopefully the test suite will pick up anything that goes wrong.

Comment on lines 1987 to 2002
/// This is the wrapper around `ty::Const` for a non-local constant. Because it doesn't have a
/// `BodyId`, we need to handle it on its own.
TyConst { expr: String },
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
/// This is the wrapper around `ty::Const` for a non-local constant. Because it doesn't have a
/// `BodyId`, we need to handle it on its own.
TyConst { expr: String },
/// This is the wrapper around `ty::Const` for a non-local constant. Because it doesn't have a
/// `BodyId`, we need to handle it on its own.
///
/// Note that `ty::Const` includes generic parameters, and may not always be uniquely identified by a DefId.
/// So this field must be different from `Extern`.
TyConst { expr: String },

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

👍

Comment on lines +90 to +105
crate trait FromWithTcx<T> {
fn from_tcx(f: T, tcx: TyCtxt<'_>) -> Self;
}

crate trait IntoWithTcx<T> {
fn into_tcx(self, tcx: TyCtxt<'_>) -> T;
}

impl<T, U> IntoWithTcx<U> for T
where
U: FromWithTcx<T>,
{
fn into_tcx(self, tcx: TyCtxt<'_>) -> U {
U::from_tcx(self, tcx)
}
}
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why do these need to be traits? It seems simpler to just use free functions. Otherwise you'll have to redo this all again next time you want to add a parameter.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I kept it as close as possible to the current code (which was using From). Otherwise I'll have to replace all the current trait implementations with free functions for all different types.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hmm, well I'm not convinced the traits are pulling their weight, but we can fix it in a follow-up to avoid unnecessary churn.

@GuillaumeGomez
Copy link
Member Author

Rebased to fix conflicts and updated doc comment as suggested by @jyn514. Now remains #82873 (comment).

@jyn514
Copy link
Member

jyn514 commented Mar 25, 2021

@bors r+

@bors
Copy link
Contributor

bors commented Mar 25, 2021

📌 Commit 7b59089 has been approved by jyn514

@bors bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Mar 25, 2021
@bors
Copy link
Contributor

bors commented Mar 25, 2021

⌛ Testing commit 7b59089 with merge 3debe9a...

@bors
Copy link
Contributor

bors commented Mar 26, 2021

☀️ Test successful - checks-actions
Approved by: jyn514
Pushing 3debe9a to master...

@bors bors added the merged-by-bors This PR was explicitly merged by bors. label Mar 26, 2021
@bors bors merged commit 3debe9a into rust-lang:master Mar 26, 2021
@rustbot rustbot added this to the 1.53.0 milestone Mar 26, 2021
@GuillaumeGomez GuillaumeGomez deleted the rustdoc-const-ty branch March 26, 2021 08:52
wip-sync pushed a commit to NetBSD/pkgsrc-wip that referenced this pull request Jun 20, 2021
Pkgsrc changes:
 * Bump bootstrap requirements to 1.52.1
 * Adjust patches, adapt to upstream changes, adjust cargo checksums
 * If using an external llvm, require >= 10.0

Upsteream changes:

Version 1.53.0 (2021-06-17)
============================

Language
-----------------------
- [You can now use unicode for identifiers.][83799] This allows
  multilingual identifiers but still doesn't allow glyphs that are
  not considered characters such as `#` (diamond) or `<U+1F980>`
  (crab). More specifically you can now use any identifier that
  matches the UAX #31 "Unicode Identifier and Pattern Syntax" standard. This
  is the same standard as languages like Python, however Rust uses NFC
  normalization which may be different from other languages.
- [You can now specify "or patterns" inside pattern matches.][79278]
  Previously you could only use `|` (OR) on complete patterns. E.g.
  ```rust
  let x = Some(2u8);
  // Before
  matches!(x, Some(1) | Some(2));
  // Now
  matches!(x, Some(1 | 2));
  ```
- [Added the `:pat_param` `macro_rules!` matcher.][83386] This matcher
  has the same semantics as the `:pat` matcher. This is to allow `:pat`
  to change semantics to being a pattern fragment in a future edition.

Compiler
-----------------------
- [Updated the minimum external LLVM version to LLVM 10.][83387]
- [Added Tier 3\* support for the `wasm64-unknown-unknown` target.][80525]
- [Improved debuginfo for closures and async functions on Windows MSVC.][83941]

\* Refer to Rust's [platform support page][platform-support-doc] for more
information on Rust's tiered platform support.

Libraries
-----------------------
- [Abort messages will now forward to `android_set_abort_message` on
  Android platforms when available.][81469]
- [`slice::IterMut<'_, T>` now implements `AsRef<[T]>`][82771]
- [Arrays of any length now implement `IntoIterator`.][84147]
  Currently calling `.into_iter()` as a method on an array will
  return `impl Iterator<Item=&T>`, but this may change in a
  future edition to change `Item` to `T`. Calling `IntoIterator::into_iter`
  directly on arrays will provide `impl Iterator<Item=T>` as expected.
- [`leading_zeros`, and `trailing_zeros` are now available on all
  `NonZero` integer types.][84082]
- [`{f32, f64}::from_str` now parse and print special values
  (`NaN`, `-0`) according to IEEE RFC 754.][78618]
- [You can now index into slices using `(Bound<usize>, Bound<usize>)`.][77704]
- [Add the `BITS` associated constant to all numeric types.][82565]

Stabilised APIs
---------------
- [`AtomicBool::fetch_update`]
- [`AtomicPtr::fetch_update`]
- [`BTreeMap::retain`]
- [`BTreeSet::retain`]
- [`BufReader::seek_relative`]
- [`DebugStruct::non_exhaustive`]
- [`Duration::MAX`]
- [`Duration::ZERO`]
- [`Duration::is_zero`]
- [`Duration::saturating_add`]
- [`Duration::saturating_mul`]
- [`Duration::saturating_sub`]
- [`ErrorKind::Unsupported`]
- [`Option::insert`]
- [`Ordering::is_eq`]
- [`Ordering::is_ge`]
- [`Ordering::is_gt`]
- [`Ordering::is_le`]
- [`Ordering::is_lt`]
- [`Ordering::is_ne`]
- [`OsStr::is_ascii`]
- [`OsStr::make_ascii_lowercase`]
- [`OsStr::make_ascii_uppercase`]
- [`OsStr::to_ascii_lowercase`]
- [`OsStr::to_ascii_uppercase`]
- [`Peekable::peek_mut`]
- [`Rc::decrement_strong_count`]
- [`Rc::increment_strong_count`]
- [`Vec::extend_from_within`]
- [`array::from_mut`]
- [`array::from_ref`]
- [`char::MAX`]
- [`char::REPLACEMENT_CHARACTER`]
- [`char::UNICODE_VERSION`]
- [`char::decode_utf16`]
- [`char::from_digit`]
- [`char::from_u32_unchecked`]
- [`char::from_u32`]
- [`cmp::max_by_key`]
- [`cmp::max_by`]
- [`cmp::min_by_key`]
- [`cmp::min_by`]
- [`f32::is_subnormal`]
- [`f64::is_subnormal`]

Cargo
-----------------------
- [Cargo now supports git repositories where the default `HEAD` branch is not
  "master".][cargo/9392] This also includes a switch to the version
  3 `Cargo.lock` format which can handle default branches correctly.
- [macOS targets now default to `unpacked` split-debuginfo.][cargo/9298]
- [The `authors` field is no longer included in `Cargo.toml` for new
  projects.][cargo/9282]

Rustdoc
-----------------------
- [Added the `rustdoc::bare_urls` lint that warns when you have URLs
  without hyperlinks.][81764]

Compatibility Notes
-------------------
- [Implement token-based handling of attributes during expansion][82608]
- [`Ipv4::from_str` will now reject octal format IP addresses in addition
  to rejecting hexadecimal IP addresses.][83652] The octal format can lead
  to confusion and potential security vulnerabilities and [is no
  longer recommended][ietf6943].

Internal Only
-------------
These changes provide no direct user facing benefits, but represent significant
improvements to the internals and overall performance of rustc and
related tools.

- [Rework the `std::sys::windows::alloc` implementation.][83065]
- [rustdoc: Don't enter an infer_ctxt in get_blanket_impls for
  impls that aren't blanket impls.][82864]
- [rustdoc: Only look at blanket impls in `get_blanket_impls`][83681]
- [Rework rustdoc const type][82873]

[83386]: rust-lang/rust#83386
[82771]: rust-lang/rust#82771
[84147]: rust-lang/rust#84147
[84082]: rust-lang/rust#84082
[83799]: rust-lang/rust#83799
[83681]: rust-lang/rust#83681
[83652]: rust-lang/rust#83652
[83387]: rust-lang/rust#83387
[82873]: rust-lang/rust#82873
[82864]: rust-lang/rust#82864
[82608]: rust-lang/rust#82608
[82565]: rust-lang/rust#82565
[80525]: rust-lang/rust#80525
[79278]: rust-lang/rust#79278
[78618]: rust-lang/rust#78618
[77704]: rust-lang/rust#77704
[83941]: rust-lang/rust#83941
[83065]: rust-lang/rust#83065
[81764]: rust-lang/rust#81764
[81469]: rust-lang/rust#81469
[cargo/9298]: rust-lang/cargo#9298
[cargo/9282]: rust-lang/cargo#9282
[cargo/9392]: rust-lang/cargo#9392
[`char::MAX`]: https://doc.rust-lang.org/std/primitive.char.html#associatedconstant.MAX
[`char::REPLACEMENT_CHARACTER`]: https://doc.rust-lang.org/std/primitive.char.html#associatedconstant.REPLACEMENT_CHARACTER
[`char::UNICODE_VERSION`]: https://doc.rust-lang.org/std/primitive.char.html#associatedconstant.UNICODE_VERSION
[`char::decode_utf16`]: https://doc.rust-lang.org/std/primitive.char.html#method.decode_utf16
[`char::from_u32`]: https://doc.rust-lang.org/std/primitive.char.html#method.from_u32
[`char::from_u32_unchecked`]: https://doc.rust-lang.org/std/primitive.char.html#method.from_u32_unchecked
[`char::from_digit`]: https://doc.rust-lang.org/std/primitive.char.html#method.from_digit
[`AtomicBool::fetch_update`]: https://doc.rust-lang.org/std/sync/atomic/struct.AtomicBool.html#method.fetch_update
[`AtomicPtr::fetch_update`]: https://doc.rust-lang.org/std/sync/atomic/struct.AtomicPtr.html#method.fetch_update
[`BTreeMap::retain`]: https://doc.rust-lang.org/std/collections/struct.BTreeMap.html#method.retain
[`BTreeSet::retain`]: https://doc.rust-lang.org/std/collections/struct.BTreeSet.html#method.retain
[`BufReader::seek_relative`]: https://doc.rust-lang.org/std/io/struct.BufReader.html#method.seek_relative
[`DebugStruct::non_exhaustive`]: https://doc.rust-lang.org/std/fmt/struct.DebugStruct.html#method.finish_non_exhaustive
[`Duration::MAX`]: https://doc.rust-lang.org/std/time/struct.Duration.html#associatedconstant.MAX
[`Duration::ZERO`]: https://doc.rust-lang.org/std/time/struct.Duration.html#associatedconstant.ZERO
[`Duration::is_zero`]: https://doc.rust-lang.org/std/time/struct.Duration.html#method.is_zero
[`Duration::saturating_add`]: https://doc.rust-lang.org/std/time/struct.Duration.html#method.saturating_add
[`Duration::saturating_mul`]: https://doc.rust-lang.org/std/time/struct.Duration.html#method.saturating_mul
[`Duration::saturating_sub`]: https://doc.rust-lang.org/std/time/struct.Duration.html#method.saturating_sub
[`ErrorKind::Unsupported`]: https://doc.rust-lang.org/std/io/enum.ErrorKind.html#variant.Unsupported
[`Option::insert`]: https://doc.rust-lang.org/std/option/enum.Option.html#method.insert
[`Ordering::is_eq`]: https://doc.rust-lang.org/std/cmp/enum.Ordering.html#method.is_eq
[`Ordering::is_ge`]: https://doc.rust-lang.org/std/cmp/enum.Ordering.html#method.is_ge
[`Ordering::is_gt`]: https://doc.rust-lang.org/std/cmp/enum.Ordering.html#method.is_gt
[`Ordering::is_le`]: https://doc.rust-lang.org/std/cmp/enum.Ordering.html#method.is_le
[`Ordering::is_lt`]: https://doc.rust-lang.org/std/cmp/enum.Ordering.html#method.is_lt
[`Ordering::is_ne`]: https://doc.rust-lang.org/std/cmp/enum.Ordering.html#method.is_ne
[`OsStr::eq_ignore_ascii_case`]: https://doc.rust-lang.org/std/ffi/struct.OsStr.html#method.eq_ignore_ascii_case
[`OsStr::is_ascii`]: https://doc.rust-lang.org/std/ffi/struct.OsStr.html#method.is_ascii
[`OsStr::make_ascii_lowercase`]: https://doc.rust-lang.org/std/ffi/struct.OsStr.html#method.make_ascii_lowercase
[`OsStr::make_ascii_uppercase`]: https://doc.rust-lang.org/std/ffi/struct.OsStr.html#method.make_ascii_uppercase
[`OsStr::to_ascii_lowercase`]: https://doc.rust-lang.org/std/ffi/struct.OsStr.html#method.to_ascii_lowercase
[`OsStr::to_ascii_uppercase`]: https://doc.rust-lang.org/std/ffi/struct.OsStr.html#method.to_ascii_uppercase
[`Peekable::peek_mut`]: https://doc.rust-lang.org/std/iter/struct.Peekable.html#method.peek_mut
[`Rc::decrement_strong_count`]: https://doc.rust-lang.org/std/rc/struct.Rc.html#method.increment_strong_count
[`Rc::increment_strong_count`]: https://doc.rust-lang.org/std/rc/struct.Rc.html#method.increment_strong_count
[`Vec::extend_from_within`]: https://doc.rust-lang.org/beta/std/vec/struct.Vec.html#method.extend_from_within
[`array::from_mut`]: https://doc.rust-lang.org/beta/std/array/fn.from_mut.html
[`array::from_ref`]: https://doc.rust-lang.org/beta/std/array/fn.from_ref.html
[`cmp::max_by_key`]: https://doc.rust-lang.org/beta/std/cmp/fn.max_by_key.html
[`cmp::max_by`]: https://doc.rust-lang.org/beta/std/cmp/fn.max_by.html
[`cmp::min_by_key`]: https://doc.rust-lang.org/beta/std/cmp/fn.min_by_key.html
[`cmp::min_by`]: https://doc.rust-lang.org/beta/std/cmp/fn.min_by.html
[`f32::is_subnormal`]: https://doc.rust-lang.org/std/primitive.f64.html#method.is_subnormal
[`f64::is_subnormal`]: https://doc.rust-lang.org/std/primitive.f64.html#method.is_subnormal
[ietf6943]: https://datatracker.ietf.org/doc/html/rfc6943#section-3.1.1
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
merged-by-bors This PR was explicitly merged by bors. relnotes-perf Performance improvements that should be mentioned in the release notes. S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. T-rustdoc Relevant to the rustdoc team, which will review and decide on the PR/issue.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

9 participants