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

tracing: Instrument std::future::Future in tracing #808

Merged
merged 16 commits into from
Sep 25, 2020

Conversation

davidbarsky
Copy link
Member

@davidbarsky davidbarsky commented Jul 12, 2020

This branch introduces tracing::Instrument, a trait borrowed from tracing-futures. Instrument is implemented on std::future::Future, allowing tracing users instrument async functions and futures with spans without needing to depend on tracing-futures. Additionally, tracing_macros::instrument macro will now just work when used with async functions.

Few notes:

  • tracing depend depends on pin-project-lite, inline with just as Tokio does.
  • Before merging this branch, I'd like to look over the docs to ensure that inaccurate statements along the lines of "use tracing-futures for std::future::Future" are removed. However, since some of these changes would refer to as-of-yet non-existent links on docs.rs, I'd like to get this shipped as quickly as possible.

@davidbarsky davidbarsky requested a review from a team July 12, 2020 01:26
@hawkw
Copy link
Member

hawkw commented Jul 12, 2020

* Should `tracing` depend on `pin-project` or `pin-project-lite`? While `tracing` compiles syn and quote by default, `tracing`'s [proposed inclusion](https://github.com/tokio-rs/tokio/pull/2655) into Tokio might become problematic.

If we don't need any of the fancier stuff from pin-project (and I don't believe we do, we just need basic field projection), I think it's better to use pin-project-lite, IMO. Remember that the syn/quote dependency can be disabled by disabling the attributes feature flag...

@davidbarsky davidbarsky marked this pull request as ready for review July 20, 2020 21:24
@davidbarsky davidbarsky requested a review from hawkw as a code owner July 20, 2020 21:24
Copy link
Member

@hawkw hawkw left a comment

Choose a reason for hiding this comment

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

This looks basically right. I commented on a few things.

Also, note that my proposal for making Span::in_scope work with async blocks as well as closures (PR #734) requires some changes to the Instrument trait:

tracing/tracing/src/span.rs

Lines 317 to 320 in 85995a5

pub trait Instrument<I: Instrumented> {
fn instrument(self, span: &Span) -> I::Output
where
Self: Sized,
If we decide to make this change, we may want to consider making the API changes in Instrument necessary for this idea to work before we publish a version of tracing with the Instrument trait, so we don't need to break it later...

tracing/src/instrument.rs Outdated Show resolved Hide resolved
tracing/src/instrument.rs Outdated Show resolved Hide resolved
tracing/src/instrument.rs Outdated Show resolved Hide resolved
tracing/src/instrument.rs Outdated Show resolved Hide resolved
tracing/src/instrument.rs Outdated Show resolved Hide resolved
Comment on lines 97 to 98
/// [`Subscriber`]: https://docs.rs/tracing/latest/tracing/subscriber/trait.Subscriber.html
/// [default]: https://docs.rs/tracing/latest/tracing/dispatcher/index.html#setting-the-default-subscriber
Copy link
Member

Choose a reason for hiding this comment

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

probably should be in-crate links

Comment on lines 120 to 121
/// [`Subscriber`]: https://docs.rs/tracing/latest/tracing/subscriber/trait.Subscriber.html
/// [default]: https://docs.rs/tracing/latest/tracing/dispatcher/index.html#setting-the-default-subscriber
Copy link
Member

Choose a reason for hiding this comment

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

probably should be in-crate links.

tracing/src/instrument.rs Outdated Show resolved Hide resolved
tracing/src/instrument.rs Outdated Show resolved Hide resolved
/// a `tracing` [`Subscriber`].
///
/// [`Subscriber`]: https://docs.rs/tracing/latest/tracing/subscriber/trait.Subscriber.html
pub trait WithSubscriber: Sized {
Copy link
Member

Choose a reason for hiding this comment

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

i'm on the fence about whether or not this needs to be in tracing proper. it isn't a whole lot of code, but I also think it's probably fine for it to stay in tracing-futures...

@davidbarsky
Copy link
Member Author

davidbarsky commented Jul 21, 2020

Resolved the in-crate links, but I'm not sure how to link to an HTML section using intra-doc links in Rustdoc.

I'm a little on the fence on the instrument trait change, but I don't feel too strongly.

@davidbarsky
Copy link
Member Author

Actually, let's do the change. Do you want to land it as part of your (rebased) PR or in this PR?

@davidbarsky davidbarsky requested a review from hawkw July 21, 2020 18:43
Comment on lines +119 to +120
.instrument(info_span!("echo", %peer_addr))
.await?;
Copy link
Member

Choose a reason for hiding this comment

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

whoopsy

Copy link
Member Author

Choose a reason for hiding this comment

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

yup 😬

@hawkw
Copy link
Member

hawkw commented Jul 22, 2020

Windows CI failure for 2c4f3d1 was definitely not your fault:

error: could not download file from 'https://static.rust-lang.org/rustup/release-stable.toml' to 'C:\Users\RUNNER~1\AppData\Local\Temp\rustup-updateMwhrLq\release-stable.toml'
error: caused by: failed to make network request
error: caused by: error sending request for url (https://static.rust-lang.org/rustup/release-stable.toml): connection error: An existing connection was forcibly closed by the remote host. (os error 10054)
error: caused by: connection error: An existing connection was forcibly closed by the remote host. (os error 10054)
error: caused by: An existing connection was forcibly closed by the remote host. (os error 10054)

I restarted it.

@hawkw
Copy link
Member

hawkw commented Jul 22, 2020

Actually, let's do the change. Do you want to land it as part of your (rebased) PR or in this PR?

Hmm.

Let's not make the change to the in_scope function as part of this PR --- I think it might be considered semver breaking, since it changes a function signature. It won't break callers of the function, since any previously-accepted input types would still be accepted, but it changes the actual signature of the function, which would potentially be an issue if it was e.g. being used as a function pointer.

We can land the version of the Instrument trait as proposed in that PR as part of this branch, since it's an additive change. If you like, I can open a PR against this branch to make those changes here?

@hawkw
Copy link
Member

hawkw commented Jul 22, 2020

Let's not make the change to the in_scope function as part of this PR --- I think it might be considered semver breaking, since it changes a function signature. It won't break callers of the function, since any previously-accepted input types would still be accepted, but it changes the actual signature of the function, which would potentially be an issue if it was e.g. being used as a function pointer.

A potential option for making the whole change without breaking is to just add a new Span method for the version of in_scope that also works for async blocks, under a new name, and (potentially) deprecating in_scope in favor of that new method. The downside of that is that we would have to think of a better name than in_scope...

@davidbarsky
Copy link
Member Author

The proposed changes feels like a breaking change due to the signature change and I'd rather be conservative with these changes. Since we can intend to ship tracing 0.2 soon (TM), I'm okay with bundling that change into that release instead of this PR.

@hawkw
Copy link
Member

hawkw commented Jul 22, 2020

The proposed changes feels like a breaking change due to the signature change and I'd rather be conservative with these changes. Since we can intend to ship tracing 0.2 soon (TM), I'm okay with bundling that change into that release instead of this PR.

Okay, that seems fine to me. Let's just make sure that we land the Instrument trait in a form that will allow us to make that change without breaking Instrument (e.g.: land the Instrument trait from that PR).

@LegNeato
Copy link

LegNeato commented Sep 3, 2020

Any update on this?

@davidbarsky davidbarsky requested a review from hawkw September 24, 2020 22:43
Comment on lines +909 to +910
/// Attach a span to a `std::future::Future`.
pub mod instrument;
Copy link
Member

Choose a reason for hiding this comment

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

IDK if I really think this needs its own module, I might either reexport it from the root and make the instrument module hidden, or put it in the span module. but, i don't really care.

@davidbarsky davidbarsky merged commit a8cc977 into master Sep 25, 2020
@davidbarsky davidbarsky deleted the davidbarsky/add-instrument-trait-to-tracing branch September 25, 2020 20:18
hawkw added a commit that referenced this pull request Sep 28, 2020
Fixed

- Incorrect inlining of `Span::new`, `Span::new_root`, and
  `Span::new_child_of`, which could result in `dispatcher::get_default`
  being  inlined at the callsite ([#994])
- Regression where using a struct field as a span or event field when
  other fields on that struct are borrowed mutably would fail to compile
  ([#987])

Changed

- Updated `tracing-core` to 0.1.17 ([#992])

Added

- `Instrument` trait and `Instrumented` type for attaching a `Span` to a
  `Future` ([#808])
- `Copy` implementations for `Level` and `LevelFilter` ([#992])
- Multiple documentation fixes and improvements ([#964], [#980], [#981])

Thanks to @nagisa, and new contributors @securityinsanity, @froydnj,
@jyn514 and @TaKO8Ki for contributing to this release!

[#994]: #994
[#992]: #992
[#987]: #987
[#980]: #980
[#981]: #981
[#964]: #964
[#808]: #808

Signed-off-by: Eliza Weisman <[email protected]>
hawkw added a commit that referenced this pull request Sep 28, 2020
### Fixed

- Incorrect inlining of `Span::new`, `Span::new_root`, and
  `Span::new_child_of`, which could result in `dispatcher::get_default`
  being  inlined at the callsite ([#994])
- Regression where using a struct field as a span or event field when
  other fields on that struct are borrowed mutably would fail to compile
  ([#987])

### Changed

- Updated `tracing-core` to 0.1.17 ([#992])

### Added

- `Instrument` trait and `Instrumented` type for attaching a `Span` to a
  `Future` ([#808])
- `Copy` implementations for `Level` and `LevelFilter` ([#992])
- Multiple documentation fixes and improvements ([#964], [#980], [#981])

Thanks to @nagisa, and new contributors @securityinsanity, @froydnj,
@jyn514 and @TaKO8Ki for contributing to this release!

[#994]: #994
[#992]: #992
[#987]: #987
[#980]: #980
[#981]: #981
[#964]: #964
[#808]: #808

Signed-off-by: Eliza Weisman <[email protected]>
dvdplm added a commit to dvdplm/tracing that referenced this pull request Oct 5, 2020
* master:
  tracing: Instrument std::future::Future (tokio-rs#808)
  tracing: move `ValueSet` construction out of closures (tokio-rs#987)
hawkw added a commit that referenced this pull request Feb 4, 2021
Fixed

- Compiler error when using `#[instrument(err)]` on functions with
  mutable parameters ([#1167])
- Missing function visibility modifier when using `#[instrument]` with
  `async-trait` ([#977])
- Multiple documentation fixes and improvements ([#965], [#981],
  [#1215])

 Changed

- `tracing-futures` dependency is no longer required when using
  `#[instrument]` on async functions ([#808])

Thanks to @nagisa, @Txuritan, @TaKO8Ki, and @okready for contributing to
this release!

[#1167]: #1167
[#977]: #977
[#965]: #965
[#981]: #981
[#1215]: #1215
[#808]: #808
hawkw added a commit that referenced this pull request Feb 4, 2021
Fixed

- Compiler error when using `#[instrument(err)]` on functions with
  mutable parameters ([#1167])
- Missing function visibility modifier when using `#[instrument]` with
  `async-trait` ([#977])
- Multiple documentation fixes and improvements ([#965], [#981],
  [#1215])

 Changed

- `tracing-futures` dependency is no longer required when using
  `#[instrument]` on async functions ([#808])

Thanks to @nagisa, @Txuritan, @TaKO8Ki, and @okready for contributing to
this release!

[#1167]: #1167
[#977]: #977
[#965]: #965
[#981]: #981
[#1215]: #1215
[#808]: #808
hawkw added a commit that referenced this pull request Feb 4, 2021
### Fixed

- Compiler error when using `#[instrument(err)]` on functions with
  mutable parameters ([#1167])
- Missing function visibility modifier when using `#[instrument]` with
  `async-trait` ([#977])
- Multiple documentation fixes and improvements ([#965], [#981],
  [#1215])

 ### Changed

- `tracing-futures` dependency is no longer required when using
  `#[instrument]` on async functions ([#808])

Thanks to @nagisa, @Txuritan, @TaKO8Ki, and @okready for contributing to
this release!

[#1167]: #1167
[#977]: #977
[#965]: #965
[#981]: #981
[#1215]: #1215
[#808]: #808
hawkw added a commit that referenced this pull request Feb 4, 2021
Fixed

- **attributes**: Compiler error when using `#[instrument(err)]` on
  functions with mutable parameters ([#1167])
- **attributes**: Missing function visibility modifier when using
  `#[instrument]` with `async-trait` ([#977])
- **attributes** Removed unused `syn` features ([#928])
- **log**: Fixed an issue where the `tracing` macros would generate code
  for events whose levels are disabled statically by the `log` crate's
  `static_max_level_XXX` features ([#1175])
- Fixed deprecations and clippy lints ([#1195])
- Several documentation fixes and improvements ([#941], [#965], [#981],
  [#1146], [#1215])

Changed

- **attributes**: `tracing-futures` dependency is no longer required
  when using `#[instrument]` on async functions ([#808])
- **attributes**: Updated `tracing-attributes` minimum dependency to
  v0.1.12 ([#1222])

Thanks to @nagisa, @Txuritan, @TaKO8Ki, @okready, and @krojew for
contributing to this release!
hawkw added a commit that referenced this pull request Feb 4, 2021
### Fixed

- **attributes**: Compiler error when using `#[instrument(err)]` on
  functions with mutable parameters ([#1167])
- **attributes**: Missing function visibility modifier when using
  `#[instrument]` with `async-trait` ([#977])
- **attributes** Removed unused `syn` features ([#928])
- **log**: Fixed an issue where the `tracing` macros would generate code
  for events whose levels are disabled statically by the `log` crate's
  `static_max_level_XXX` features ([#1175])
- Fixed deprecations and clippy lints ([#1195])
- Several documentation fixes and improvements ([#941], [#965], [#981],
  [#1146], [#1215])

### Changed

- **attributes**: `tracing-futures` dependency is no longer required
  when using `#[instrument]` on async functions ([#808])
- **attributes**: Updated `tracing-attributes` minimum dependency to
  v0.1.12 ([#1222])

Thanks to @nagisa, @Txuritan, @TaKO8Ki, @okready, and @krojew for
contributing to this release!
kaffarell pushed a commit to kaffarell/tracing that referenced this pull request May 22, 2024
### Fixed

- Compiler error when using `#[instrument(err)]` on functions with
  mutable parameters ([tokio-rs#1167])
- Missing function visibility modifier when using `#[instrument]` with
  `async-trait` ([tokio-rs#977])
- Multiple documentation fixes and improvements ([tokio-rs#965], [tokio-rs#981],
  [tokio-rs#1215])

 ### Changed

- `tracing-futures` dependency is no longer required when using
  `#[instrument]` on async functions ([tokio-rs#808])

Thanks to @nagisa, @Txuritan, @TaKO8Ki, and @okready for contributing to
this release!

[tokio-rs#1167]: tokio-rs#1167
[tokio-rs#977]: tokio-rs#977
[tokio-rs#965]: tokio-rs#965
[tokio-rs#981]: tokio-rs#981
[tokio-rs#1215]: tokio-rs#1215
[tokio-rs#808]: tokio-rs#808
kaffarell pushed a commit to kaffarell/tracing that referenced this pull request May 22, 2024
### Fixed

- **attributes**: Compiler error when using `#[instrument(err)]` on
  functions with mutable parameters ([tokio-rs#1167])
- **attributes**: Missing function visibility modifier when using
  `#[instrument]` with `async-trait` ([tokio-rs#977])
- **attributes** Removed unused `syn` features ([tokio-rs#928])
- **log**: Fixed an issue where the `tracing` macros would generate code
  for events whose levels are disabled statically by the `log` crate's
  `static_max_level_XXX` features ([tokio-rs#1175])
- Fixed deprecations and clippy lints ([tokio-rs#1195])
- Several documentation fixes and improvements ([tokio-rs#941], [tokio-rs#965], [tokio-rs#981],
  [tokio-rs#1146], [tokio-rs#1215])

### Changed

- **attributes**: `tracing-futures` dependency is no longer required
  when using `#[instrument]` on async functions ([tokio-rs#808])
- **attributes**: Updated `tracing-attributes` minimum dependency to
  v0.1.12 ([tokio-rs#1222])

Thanks to @nagisa, @Txuritan, @TaKO8Ki, @okready, and @krojew for
contributing to this release!
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants