`rustc_builtin_macros` cleanup, part 8 by cyrgani · Pull Request #163485 · rust-lang/rust

cyrgani · 2026-09-29T12:47:02Z

This PR is roughly split into four parts:

The first few commits add helpers to ExtCtxt to both construct some common AST elements in an easier way. They also change some Ident arguments to Symbol to avoid the annoying call-site pattern of repeating the span (cx.some_method(span, Ident::new(name, span), ...)).
The next few commits focus on the derive infrastructure cleanup a bit.
The further following commits make Path construction faster and remove the new_path remnant of the old local path version.
The last four commits are two changes where I am not sure which option is nicer:
- We can either remove path_std! and pathvec! entirely or utilize them in the whole crate. Both solutions seem better than the current mixed situation.
- I added the old Substructure back in use the type name instead of Self in constructor paths again #163863 to fix the regression without thinking much about whether it is actually needed. We can either remove it again or use it to also store the &Item which the derive is applied to, which allows removing combine_substructure, but reduces the flexibility a bit.

Previous PR: #163389

r? @nnethercote

jieyouxu · 2026-09-29T13:01:32Z

@bors try @rust-timer queue

[PERF] `rustc_builtin_macros` cleanup, part 8

rust-bors · 2026-09-29T14:04:44Z

☀️ Try build successful (CI)
Build commit: 0bcf40d (0bcf40dc0c6b1a2253f428ebaa94e0e54aa98617)
Base parent: b46c90c (b46c90c5619a16a0c7e4d0f4e9e1ad4db4cf106e)

rust-timer · 2026-09-29T16:15:34Z

Finished benchmarking commit (0bcf40d): comparison URL.

Overall result: ✅ improvements - no action needed

Benchmarking means the PR may be perf-sensitive. Consider adding rollup=never if this change is not fit for rolling up.

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

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.2%	[-0.3%, -0.2%]	7
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (primary 0.7%, secondary 2.1%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	2.2%	[2.2%, 2.2%]	1
Regressions ❌ (secondary)	3.6%	[3.4%, 3.8%]	3
Improvements ✅ (primary)	-0.9%	[-0.9%, -0.9%]	1
Improvements ✅ (secondary)	-2.3%	[-2.3%, -2.3%]	1
All ❌✅ (primary)	0.7%	[-0.9%, 2.2%]	2

Cycles

This perf run didn't have relevant results for this metric.

Binary size

This perf run didn't have relevant results for this metric.

Bootstrap: 491.711s -> 491.207s (-0.10%)
Artifact size: 406.38 MiB -> 406.39 MiB (0.00%)

These are especially useful for constructing types and paths referring to primitives.

It is only used in one place and it's easier to do the check directly in `debug.rs`.

Several small changes to remove some unhelpful helpers.

As I understand it, this would only be needed if the crate defining the derive macros (which is `core`) was using edition 2015. Since `core` uses edition 2024, there is in this case no difference between global and local and we can skip the leading `::`.

Previously, `std_path` converted the given `&[Symbol]` to a `Vec<Ident>` by prepending a `$crate::` and mapping each `Symbol` to an `Ident`. This commit changes the functions that previously took `Vec<Ident>` to directly take `Path` instead, and updates `std_path` to also directly return a `Path`. This allows directly prepending the `$crate::` to the resulting `segments`. It also allows constructing a `Path` from `&[Ident]` instead of `Vec<Ident>`.

`new_path` is just a obsolete trivial wrapper around `cx.std_path_all` now.

TODO: is this actually worth it? maybe we should remove the macros entirely instead?

Passing around `type_ident` everywhere is what the old code did as well, but it's actually only needed in two places.

This reverts commit c28f2fece8497fb2119e17a47188c6ab1dc5da7b. (This revert only exists to make the diff for the next commit clearer and will be removed before merging.)

This leads to `CombineSubstructureFunc` becoming a `fn`, allows some cleanups and requires restructuring `From` a bit. It also means that it is no longer easily possible to generate the dummy method body after encountering an error, so the code just skips generating anything.

rustbot · 2026-10-09T12:24:02Z

Some changes occurred in compiler/rustc_builtin_macros/src/autodiff.rs

cc @ZuseZ4

cyrgani · 2026-10-09T13:07:28Z

 use crate::{deriving, diagnostics};

-pub(crate) struct TraitDef<'a> {
+pub(crate) struct TraitDef<> {


oh, it looks like rustfmt skips formatting a struct with a default field entirely

View changes since the review

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Sep 29, 2026

This comment has been minimized.

Sign in to view

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Sep 29, 2026

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Sep 29, 2026

Auto merge of #163485 - cyrgani:builtin-macros-8, r=<try>

0bcf40d

[PERF] `rustc_builtin_macros` cleanup, part 8

petrochenkov reviewed Sep 29, 2026

View reviewed changes

Comment thread compiler/rustc_expand/src/base.rs Outdated

This comment has been minimized.

Sign in to view

rustbot removed the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Sep 29, 2026

This comment has been minimized.

Sign in to view

cyrgani added 17 commits October 9, 2026 08:33

rename Legacy* macro kinds to Ast*

2ca3727

add cx.ty_sym and cx.path_sym

023608f

These are especially useful for constructing types and paths referring to primitives.

cx.param only needs a Symbol

79f6e3b

cx.pat_ident only needs a Symbol

928f03e

uplift call_intrinsic to cx.expr_call_intrinsic

6f7d102

add cx.item_fn

ec8ae84

add cx.item_const_underscore

aa58eeb

use even more helpers

67aa421

use feature(default_field_values) for TraitDef

a12ceed

remove AllFieldlessEnum

2022888

It is only used in one place and it's easier to do the check directly in `debug.rs`.

inline some code

d5726d3

Several small changes to remove some unhelpful helpers.

remove new_path

6c946d3

`new_path` is just a obsolete trivial wrapper around `cx.std_path_all` now.

utilize pathvec! and path_std! in the entire crate

785039a

TODO: is this actually worth it? maybe we should remove the macros entirely instead?

do the Self constructor fix properly

801f25a

Passing around `type_ident` everywhere is what the old code did as well, but it's actually only needed in two places.

Revert "do the Self constructor fix properly"

3bc62af

This reverts commit c28f2fece8497fb2119e17a47188c6ab1dc5da7b. (This revert only exists to make the diff for the next commit clearer and will be removed before merging.)

rust-cloud-vms Bot force-pushed the builtin-macros-8 branch from 7eace55 to c96f45e Compare October 9, 2026 12:08

rustbot assigned nnethercote Oct 9, 2026

cyrgani changed the title ~~[PERF] rustc_builtin_macros cleanup, part 8~~ rustc_builtin_macros cleanup, part 8 Oct 9, 2026

cyrgani marked this pull request as ready for review October 9, 2026 12:23

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Oct 9, 2026

cyrgani commented Oct 9, 2026

View reviewed changes

Navigation Menu

Repository navigation

`rustc_builtin_macros` cleanup, part 8 - #163485

`rustc_builtin_macros` cleanup, part 8#163485
cyrgani wants to merge 18 commits into
rust-lang:mainfrom
cyrgani:builtin-macros-8

cyrgani commented Sep 29, 2026 •

edited

Loading

Uh oh!

jieyouxu commented Sep 29, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

Uh oh!

rust-bors Bot commented Sep 29, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Sep 29, 2026

Uh oh!

This comment has been minimized.

rustbot commented Oct 9, 2026

Uh oh!

cyrgani Oct 9, 2026 •

edited by rustbot

Loading

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

Conversation

cyrgani commented Sep 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jieyouxu commented Sep 29, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

Uh oh!

rust-bors Bot commented Sep 29, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Sep 29, 2026

Overall result: ✅ improvements - no action needed

Instruction count

Max RSS (memory usage)

Cycles

Binary size

Uh oh!

This comment has been minimized.

rustbot commented Oct 9, 2026

Uh oh!

cyrgani Oct 9, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

cyrgani commented Sep 29, 2026 •

edited

Loading

cyrgani Oct 9, 2026 •

edited by rustbot

Loading