mGCA: Support directly represented negated literals #152139

khyperia · 2026-02-04T20:05:33Z

PatExprKind already awkwardly tacks on a negated: bool for the same purpose:

Lines 1954 to 1959 in 8bccf12

    
           Lit { 
        
               lit: Lit, 
        
               // FIXME: move this into `Lit` and handle negated literal expressions 
        
               // once instead of matching on unop neg expressions everywhere. 
        
               negated: bool, 
        
           },

perhaps one day we should indeed do that FIXME...

r? @BoxyUwU

rustbot · 2026-02-04T20:05:37Z

Some changes occurred in src/tools/clippy

cc @rust-lang/clippy

HIR ty lowering was modified

cc @fmease

BoxyUwU · 2026-02-05T12:20:09Z

can you add tests for "nonsensical" negated literals, e.g. -true { -true } -":3" { ":3" }, ah and I guess also -1_usize and { -1_usize }.

khyperia · 2026-02-06T13:03:01Z

can you add tests for "nonsensical" negated literals, e.g. -true { -true } -":3" { ":3" }, ah and I guess also -1_usize and { -1_usize }.

aha, finally stumbled across #152001 - I can't add tests for things like -1_usize until that PR (or another fix for #151625) is merged, because the place an error should be reported for -1_usize is inside that PR's code.

khyperia · 2026-02-06T13:06:58Z

perhaps one day we should indeed do that FIXME...

discussed in the Zulip: https://rust-lang.zulipchat.com/#narrow/channel/131828-t-compiler/topic/Move.20.60negated.3A.20bool.60.20into.20Lit/with/572102179 relevant quote:

oli: Unfortunately, while I still think this would be better, my previous attempt did not garner much support and some opposition. rust-lang/compiler-team#843
oli: If you want you can just nuke the FIXME in your PR entirely

khyperia · 2026-02-06T13:12:04Z

compiler/rustc_ast_lowering/src/lib.rs

+                ConstArg {
+                    hir_id: self.lower_node_id(expr.id),
+                    kind: hir::ConstArgKind::Literal { lit: literal.node, negated: true },
                    span,


note to reviewers: in addition to nuking the FIXME in my last force-push, I changed the span for this ConstArg to include the negative sign in the span (before the push, it was excluded, just the inner_expr's span). Realized it was excluded when trying to add errors for -1_usize.

(the diff in the force push is rather unclear what's happening so I figured I'd point it out)

BoxyUwU · 2026-02-06T14:14:25Z

I can't add tests for things like -1_usize until that PR is merged

oh, right yeah :3

reddevilmidzy · 2026-02-06T15:39:20Z

can you add tests for "nonsensical" negated literals, e.g. -true { -true } -":3" { ":3" }, ah and I guess also -1_usize and { -1_usize }.

aha, finally stumbled across #152001 - I can't add tests for things like -1_usize until that PR (or another fix for #151625) is merged, because the place an error should be reported for -1_usize is inside that PR's code.

Oh, sorry for the delay. I'll work on it over the weekend.

BoxyUwU · 2026-02-06T16:36:50Z

@bors r+

I'm happy to land this with it being a little broken we can always fix it going forwards :)

rust-bors · 2026-02-06T16:36:53Z

📌 Commit 54a9be4 has been approved by BoxyUwU

It is now in the queue for this repository.

…uwer Rollup of 3 pull requests Successful merges: - #152129 (MGCA: require #[type_const] on free consts too) - #152139 (mGCA: Support directly represented negated literals) - #152189 (Convert to inline diagnostics in `rustc_passes`)

@BoxyUwU

Rollup merge of #152139 - khyperia:mgca-negative-literals, r=BoxyUwU mGCA: Support directly represented negated literals fixes #152123 PatExprKind already awkwardly tacks on a `negated: bool` for the same purpose: https://github.com/rust-lang/rust/blob/8bccf1224deab49b54694c9090e577bfe90a94e6/compiler/rustc_hir/src/hir.rs#L1954-L1959 perhaps one day we should indeed do that FIXME... r? @BoxyUwU

…=oli-obk Dont strip const blocks in array lengths r? oli-obk mGCA now handles const blocks by *always* handling them during `lower_expr_to_const_arg_direct` instead of *sometimes* stripping them out at parse time. This is just generally a lot clearer/nicer but also means parsing isn't lossy which is just straight up wrong. We now use `MgcaDisambiguation::Direct` for const blocks because we "directly" represent a const block as `hir::ConstArgKind::Anon` :> The only time that an anon const for const generics uses `MgcaDisambiguation::AnonConst` is for unbraced literals. Once we properly support literals in `hir::ConstArgKind` (see rust-lang#152139 rust-lang#152001) then `MgcaDisambiguation` can be renamed to `AnonConstKind` with `TypeSystem` and `NonTypeSystem` variants. We can also get rid of `mgca_direct_lit_hack`. I expect this to be a very nice cleanup :) Fixes rust-lang/rustfmt#6788 The diff relating to passing spans around is to avoid a bunch of mGCA diagnostics changing from `const {}` to `{}`. I'm not entirely sure why this was happening. cc @rust-lang/rustfmt How do I run the tests in the rustfmt repo from here? `x test rustfmt` only seems to run like 100 tests and doesn't result in a `target/issue-6788.rs` getting created. I've verified locally that this formats correctly though

Rollup merge of #152234 - BoxyUwU:dont_strip_const_blocks, r=oli-obk Dont strip const blocks in array lengths r? oli-obk mGCA now handles const blocks by *always* handling them during `lower_expr_to_const_arg_direct` instead of *sometimes* stripping them out at parse time. This is just generally a lot clearer/nicer but also means parsing isn't lossy which is just straight up wrong. We now use `MgcaDisambiguation::Direct` for const blocks because we "directly" represent a const block as `hir::ConstArgKind::Anon` :> The only time that an anon const for const generics uses `MgcaDisambiguation::AnonConst` is for unbraced literals. Once we properly support literals in `hir::ConstArgKind` (see #152139 #152001) then `MgcaDisambiguation` can be renamed to `AnonConstKind` with `TypeSystem` and `NonTypeSystem` variants. We can also get rid of `mgca_direct_lit_hack`. I expect this to be a very nice cleanup :) Fixes rust-lang/rustfmt#6788 The diff relating to passing spans around is to avoid a bunch of mGCA diagnostics changing from `const {}` to `{}`. I'm not entirely sure why this was happening. cc @rust-lang/rustfmt How do I run the tests in the rustfmt repo from here? `x test rustfmt` only seems to run like 100 tests and doesn't result in a `target/issue-6788.rs` getting created. I've verified locally that this formats correctly though

rustbot assigned BoxyUwU Feb 4, 2026

mGCA: Support directly represented negated literals

54a9be4

khyperia force-pushed the mgca-negative-literals branch from 7969389 to 54a9be4 Compare February 6, 2026 13:07

khyperia commented Feb 6, 2026

View reviewed changes

BoxyUwU mentioned this pull request Feb 6, 2026

Dont strip const blocks in array lengths #152234

Merged

rust-bors bot 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 Feb 6, 2026

JonathanBrouwer mentioned this pull request Feb 6, 2026

Rollup of 3 pull requests #152239

Merged

BoxyUwU mentioned this pull request Feb 6, 2026

mGCA: remove parser special case for literals #152246

Open

rust-bors bot merged commit 30bbeae into rust-lang:main Feb 6, 2026
11 checks passed

rustbot added this to the 1.95.0 milestone Feb 6, 2026

khyperia deleted the mgca-negative-literals branch February 7, 2026 09:44

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

mGCA: Support directly represented negated literals #152139

mGCA: Support directly represented negated literals #152139

Uh oh!

khyperia commented Feb 4, 2026 •

edited by rustbot

Loading

Uh oh!

rustbot commented Feb 4, 2026

Uh oh!

BoxyUwU commented Feb 5, 2026 •

edited

Loading

Uh oh!

khyperia commented Feb 6, 2026

Uh oh!

khyperia commented Feb 6, 2026

Uh oh!

khyperia Feb 6, 2026

Uh oh!

BoxyUwU commented Feb 6, 2026

Uh oh!

reddevilmidzy commented Feb 6, 2026

Uh oh!

BoxyUwU commented Feb 6, 2026

Uh oh!

rust-bors bot commented Feb 6, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

	Lit {
	lit: Lit,
	// FIXME: move this into `Lit` and handle negated literal expressions
	// once instead of matching on unop neg expressions everywhere.
	negated: bool,
	},

Uh oh!

mGCA: Support directly represented negated literals #152139

mGCA: Support directly represented negated literals #152139

Uh oh!

Conversation

khyperia commented Feb 4, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Feb 4, 2026

Uh oh!

BoxyUwU commented Feb 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

khyperia commented Feb 6, 2026

Uh oh!

khyperia commented Feb 6, 2026

Uh oh!

khyperia Feb 6, 2026

Choose a reason for hiding this comment

Uh oh!

BoxyUwU commented Feb 6, 2026

Uh oh!

reddevilmidzy commented Feb 6, 2026

Uh oh!

BoxyUwU commented Feb 6, 2026

Uh oh!

rust-bors bot commented Feb 6, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

khyperia commented Feb 4, 2026 •

edited by rustbot

Loading

BoxyUwU commented Feb 5, 2026 •

edited

Loading