Resolve lang items in AST #103662

camsteffen · 2022-10-28T00:45:24Z

Lang items can be resolved in the AST rather than the HIR. And then any HIR structures referencing LangItem can be replaced with their "normal" couterpart. Specifically QPath::LangItem and GenericBound::LangItemTrait are factored out.

Best reviewed commit-at-a-time.

rustbot · 2022-10-28T00:45:30Z

r? @TaKO8Ki

(rustbot has picked a reviewer for you, use r? to override)

rustbot · 2022-10-28T00:45:34Z

rustc_error_messages was changed

cc @davidtwco, @compiler-errors, @JohnTitor, @estebank, @TaKO8Ki

Some changes occurred in src/tools/clippy

cc @rust-lang/clippy

Some changes occurred in need_type_info.rs

cc @lcnr

camsteffen · 2022-10-31T21:52:35Z

Changed the self-type path segment to a dummy segment (Res::Err) instead of using the parent lang item.

compiler/rustc_resolve/src/lib.rs

rustbot · 2022-11-02T15:47:57Z

Hey! It looks like you've submitted a new PR for the library teams!

If this PR contains changes to any rust-lang/rust public library APIs then please comment with @rustbot label +T-libs-api -T-libs to tag it appropriately. If this PR contains changes to any unstable APIs please edit the PR description to add a link to the relevant API Change Proposal or create one if you haven't already. If you're unsure where your change falls no worries, just leave it as is and the reviewer will take a look and make a decision to forward on if necessary.

Examples of T-libs-api changes:

Stabilizing library features
Introducing insta-stable changes such as new implementations of existing stable traits on existing stable types
Introducing new or changing existing unstable library APIs (excluding permanently unstable features / features without a tracking issue)
Changing public documentation in ways that create new stability guarantees
Changing observable runtime behavior of library APIs

camsteffen · 2022-11-02T16:31:58Z

@rustbot ready (sorry for the false alarm earlier)

bors · 2022-11-02T21:40:53Z

☔ The latest upstream changes (presumably #103888) made this pull request unmergeable. Please resolve the merge conflicts.

petrochenkov · 2022-11-01T16:51:48Z

compiler/rustc_ast/src/ast.rs

+
+    pub fn generics(&self) -> Option<&Generics> {
+        match self {
+            Self::Const(..) => None,


Suggested change

Self::Const(..) => None,

AssocItemKind::Const(..) => None,

Could you avoid using Self outside of generic contexts, it makes it harder to search and read code.

petrochenkov · 2022-11-01T16:53:59Z

compiler/rustc_resolve/src/lang_items.rs

@@ -0,0 +1,126 @@
+use crate::errors::{
+    IncorrectTarget, LangItemOnIncorrectTarget, UnknownExternLangItem, UnknownLangItem,
+};


Nit: could you split imports so they fit on a single line?

petrochenkov · 2022-11-01T17:00:45Z

compiler/rustc_resolve/src/lang_items.rs

+        visit::walk_crate(&mut collector, krate);
+        let LanguageItemCollector { items, missing, .. } = collector;
+        self.lang_items = items;
+        self.missing_lang_items = missing;


You could keep a mutable reference to Resolver and push directly to resolver.items/resolver.missing instead of keeping temporary copies.

petrochenkov · 2022-11-01T17:03:20Z

compiler/rustc_resolve/src/lib.rs

@@ -955,6 +957,8 @@ pub struct Resolver<'a> {
    /// the surface (`macro` items in libcore), but are actually attributes or derives.
    builtin_macro_kinds: FxHashMap<LocalDefId, MacroKind>,
    registered_tools: RegisteredTools,
+    lang_items: Vec<(LocalDefId, LangItem)>,
+    missing_lang_items: Vec<LangItem>,


Nit: could you move it to some other place (e.g. to the end), so these fields don't end up in the middle of macro-related stuff.

petrochenkov · 2022-11-04T08:27:07Z

compiler/rustc_hir/src/hir.rs

-    }
+    // range literals desugar to a struct literal or `RangeInclusive::new(..)`
+    matches!(expr.kind, ExprKind::Call(..) | ExprKind::Struct(..))
+        && expr.span.is_desugaring(DesugaringKind::RangeLiteral)


I think it should be possible to use comparisons with lang item def ids to detect range expressions without introducing a new DesugaringKind.
Using expansion machinery in AST lowering is not a good practice, and I'd like to try eliminating it somehow.

I've wondered about this too. Do you think all of hygiene::ExpnKind::{AstPass, Desugaring, Inlined} are bad?

Tentatively yes, I just never had time to think what better alternatives are possible.
(Except for AstPass, it's used just like it's supposed to be.)

Maybe it could work similar to hir_attrs.

petrochenkov · 2022-11-04T08:34:27Z

library/std/src/panicking.rs

@@ -518,7 +518,7 @@ pub fn panicking() -> bool {
 }

 /// Entry point of panics from the libcore crate (`panic_impl` lang item).
-#[cfg(not(test))]
+#[cfg(not(any(test, doctest)))]


I guess this is necessary because rustdoc unconditionally runs the lang item collection pass now.
We probably need to run benchmarks on this PR to check how it affects performance.

petrochenkov · 2022-11-04T08:43:16Z

src/test/ui/async-await/issue-98634.rs

@@ -45,6 +45,5 @@ fn main() {
        StructAsync { callback }.await;
        //~^ ERROR expected `fn() -> impl Future<Output = ()> {callback}` to be a fn item that returns `Pin<Box<(dyn Future<Output = ()> + 'static)>>`, but it returns `impl Future<Output = ()>`
        //~| ERROR expected `fn() -> impl Future<Output = ()> {callback}` to be a fn item that returns `Pin<Box<(dyn Future<Output = ()> + 'static)>>`, but it returns `impl Future<Output = ()>`
-        //~| ERROR expected `fn() -> impl Future<Output = ()> {callback}` to be a fn item that returns `Pin<Box<(dyn Future<Output = ()> + 'static)>>`, but it returns `impl Future<Output = ()>`


It would be better if this PR was a pure refactoring not changing any logic, including diagnostics, if possible.
This PR is large and it's hard to notice and review any such changes.

petrochenkov · 2022-11-04T08:45:43Z

src/test/ui/duplicate_entry_error.stderr

@@ -1,8 +1,11 @@
 error[E0152]: found duplicate lang item `panic_impl`
  --> $DIR/duplicate_entry_error.rs:11:1
   |
-LL | fn panic_impl(info: &PanicInfo) -> ! {
-   | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+LL | / fn panic_impl(info: &PanicInfo) -> ! {


Is the "head span" logic implemented for HIR only?
It should be portable to AST pretty easily, IIRC.

petrochenkov · 2022-11-04T08:51:50Z

The HIR paths for lang items are still artificial, but this still looks like an improvement because they mimic regular paths now and can be processed by the same logic.

Was this idea discussed anywhere previously? Does it interfere with any other plans?
I remember someone mentioned moving lang item collection to an earlier stage for some reason, but I don't remember who exactly.
cc @rust-lang/compiler (I'm not going to nominate this or start an FCP, but will still ping people just in case.)

Dylan-DPC · 2022-12-15T13:46:13Z

@camsteffen any updates on this? it has bitrotted severely :P

camsteffen · 2022-12-17T16:51:58Z

We currently use QPath::LangItem to know whether a node is compiler-generated in the case of range literals. This PR currently changes that to use DesugaringKind instead. @petrochenkov noted that this is furthering a bad pattern of misusing the expansion machinery. So if we don't want to do that, then this PR is blocked on changing DesugaringKind to something better. I've experimented in that direction a bit but I don't know if I'll have a presentable solution any time soon.

Maybe this PR is a net-positive change as is (merge it), and the DesugaringKind problem can be handled down the road, but I'm not so sure. Seems just as well to let this sit. If you agree, I could write up an issue to block this on.

Dylan-DPC · 2022-12-17T16:57:46Z

If you agree, I could write up an issue to block this on.

Yeah that would be helpful.

Thanks for the update

camsteffen · 2022-12-17T22:36:40Z

As I try to put the issue into words, maybe I don't really understand what is the problem with ExpnKind::Desugaring... Desugaring and macro expansion are two very similar things (some syntax expands into different code) and it works neatly for them to share an abstraction. I do think it would be good for desugaring data to be linked to IR nodes rather than spans, but I think that is just as much true for macro expansion data.

camsteffen · 2023-03-09T22:25:22Z

Closing (for now) since I haven't had time for rust for a long while 😢

rustbot assigned TaKO8Ki Oct 28, 2022

This comment has been minimized.

Sign in to view

petrochenkov self-assigned this Oct 28, 2022

petrochenkov added S-blocked Status: Marked as blocked ❌ on something else such as an RFC or other implementation work. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Oct 28, 2022

This comment was marked as resolved.

Sign in to view

camsteffen force-pushed the lang-ast branch from bd6af71 to 79c3e85 Compare October 31, 2022 21:01

camsteffen added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-blocked Status: Marked as blocked ❌ on something else such as an RFC or other implementation work. labels Oct 31, 2022

camsteffen requested a review from petrochenkov October 31, 2022 21:07

This comment has been minimized.

Sign in to view

camsteffen force-pushed the lang-ast branch from 08605c6 to d03d743 Compare October 31, 2022 21:45

This comment has been minimized.

Sign in to view

camsteffen force-pushed the lang-ast branch from d03d743 to 7f060b0 Compare October 31, 2022 22:11

This comment has been minimized.

Sign in to view

camsteffen force-pushed the lang-ast branch from 7f060b0 to c9bfa17 Compare November 1, 2022 00:05

This comment was marked as resolved.

Sign in to view

camsteffen force-pushed the lang-ast branch 2 times, most recently from 9d50f27 to 141d06a Compare November 1, 2022 18:37

camsteffen commented Nov 1, 2022

View reviewed changes

compiler/rustc_resolve/src/lib.rs Outdated Show resolved Hide resolved

This comment has been minimized.

Sign in to view

camsteffen force-pushed the lang-ast branch from 141d06a to 409a7ac Compare November 1, 2022 19:22

This comment has been minimized.

Sign in to view

TaKO8Ki added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Nov 2, 2022

camsteffen added 7 commits November 2, 2022 09:49

Resolve lang items in ast

9e13cd3

Use inner ctxt when lowering ExprKind::Paren

5438f58

Factor out QPath::LangItem

58b5e22

Factor out GenericBound::LangItemTrait

9f480fa

Bless tests

16bfead

Fix clippy after lang item refactor

6572dc1

Cfg remove lang items in doctest

ec485f3

camsteffen force-pushed the lang-ast branch from 409a7ac to ec485f3 Compare November 2, 2022 15:47

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 Nov 2, 2022

petrochenkov reviewed Nov 4, 2022

View reviewed changes

petrochenkov added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Nov 4, 2022

Dylan-DPC added S-blocked Status: Marked as blocked ❌ on something else such as an RFC or other implementation work. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Dec 17, 2022

camsteffen closed this Mar 9, 2023

petrochenkov mentioned this pull request Nov 28, 2023

Collect lang items from AST, get rid of GenericBound::LangItemTrait #118396

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Resolve lang items in AST #103662

Resolve lang items in AST #103662

camsteffen commented Oct 28, 2022

rustbot commented Oct 28, 2022

rustbot commented Oct 28, 2022

This comment has been minimized.

This comment was marked as resolved.

This comment has been minimized.

camsteffen commented Oct 31, 2022

This comment has been minimized.

This comment has been minimized.

This comment was marked as resolved.

This comment was marked as resolved.

This comment has been minimized.

This comment has been minimized.

rustbot commented Nov 2, 2022

camsteffen commented Nov 2, 2022

bors commented Nov 2, 2022

petrochenkov Nov 1, 2022

petrochenkov Nov 1, 2022

petrochenkov Nov 1, 2022

petrochenkov Nov 1, 2022

petrochenkov Nov 4, 2022

camsteffen Nov 4, 2022

petrochenkov Nov 4, 2022

camsteffen Nov 4, 2022

petrochenkov Nov 4, 2022

petrochenkov Nov 4, 2022

petrochenkov Nov 4, 2022

petrochenkov commented Nov 4, 2022 •

edited

Loading

Dylan-DPC commented Dec 15, 2022

camsteffen commented Dec 17, 2022

Dylan-DPC commented Dec 17, 2022

camsteffen commented Dec 17, 2022

camsteffen commented Mar 9, 2023

Resolve lang items in AST #103662

Resolve lang items in AST #103662

Conversation

camsteffen commented Oct 28, 2022

rustbot commented Oct 28, 2022

rustbot commented Oct 28, 2022

This comment has been minimized.

This comment was marked as resolved.

This comment has been minimized.

camsteffen commented Oct 31, 2022

This comment has been minimized.

This comment has been minimized.

This comment was marked as resolved.

This comment was marked as resolved.

This comment has been minimized.

This comment has been minimized.

rustbot commented Nov 2, 2022

camsteffen commented Nov 2, 2022

bors commented Nov 2, 2022

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

petrochenkov commented Nov 4, 2022 • edited Loading

Dylan-DPC commented Dec 15, 2022

camsteffen commented Dec 17, 2022

Dylan-DPC commented Dec 17, 2022

camsteffen commented Dec 17, 2022

camsteffen commented Mar 9, 2023

petrochenkov commented Nov 4, 2022 •

edited

Loading