added error for invalid char cast

Author: Kivooeo

compiler/rustc_lint/messages.ftl

@@ -440,6 +440,11 @@ lint_invalid_asm_label_named = avoid using named labels in inline assembly
     .help = only local labels of the form `<number>:` should be used in inline asm
     .note = see the asm section of Rust By Example <https://doc.rust-lang.org/nightly/rust-by-example/unsafe/asm.html#labels> for more information
 lint_invalid_asm_label_no_span = the label may be declared in the expansion of a macro
+
+lint_invalid_char_cast = cannot cast `{$literal}` to `char`
+    .note = char must be a valid Unicode scalar value (code points in ranges [0x0, 0xD7FF] or [0xE000, 0x10FFFF]).
+    .note1 = The gap [0xD800, 0xDFFF] is reserved for UTF-16 surrogates and is invalid.
+
 lint_invalid_crate_type_value = invalid `crate_type` value
     .suggestion = did you mean
 

compiler/rustc_lint/src/lints.rs

@@ -1746,6 +1746,14 @@ pub(crate) struct OverflowingLiteral<'a> {
     pub lit: String,
 }
 
+#[derive(LintDiagnostic)]
+#[diag(lint_invalid_char_cast)]
+#[note]
+#[note(lint_note1)]
+pub(crate) struct InvalidCharCast {
+    pub literal: u128,
+}
+
 #[derive(LintDiagnostic)]
 #[diag(lint_uses_power_alignment)]
 pub(crate) struct UsesPowerAlignment;

compiler/rustc_lint/src/types/literal.rs

@@ -10,9 +10,9 @@ use {rustc_ast as ast, rustc_attr_data_structures as attrs, rustc_hir as hir};
 use crate::LateContext;
 use crate::context::LintContext;
 use crate::lints::{
-    OnlyCastu8ToChar, OverflowingBinHex, OverflowingBinHexSign, OverflowingBinHexSignBitSub,
-    OverflowingBinHexSub, OverflowingInt, OverflowingIntHelp, OverflowingLiteral, OverflowingUInt,
-    RangeEndpointOutOfRange, UseInclusiveRange,
+    InvalidCharCast, OnlyCastu8ToChar, OverflowingBinHex, OverflowingBinHexSign,
+    OverflowingBinHexSignBitSub, OverflowingBinHexSub, OverflowingInt, OverflowingIntHelp,
+    OverflowingLiteral, OverflowingUInt, RangeEndpointOutOfRange, UseInclusiveRange,
 };
 use crate::types::{OVERFLOWING_LITERALS, TypeLimits};
 
@@ -38,12 +38,18 @@ fn lint_overflowing_range_endpoint<'tcx>(
 
     // We only want to handle exclusive (`..`) ranges,
     // which are represented as `ExprKind::Struct`.
-    let Node::ExprField(field) = cx.tcx.parent_hir_node(hir_id) else { return false };
-    let Node::Expr(struct_expr) = cx.tcx.parent_hir_node(field.hir_id) else { return false };
+    let Node::ExprField(field) = cx.tcx.parent_hir_node(hir_id) else {
+        return false;
+    };
+    let Node::Expr(struct_expr) = cx.tcx.parent_hir_node(field.hir_id) else {
+        return false;
+    };
     if !is_range_literal(struct_expr) {
         return false;
     };
-    let ExprKind::Struct(_, [start, end], _) = &struct_expr.kind else { return false };
+    let ExprKind::Struct(_, [start, end], _) = &struct_expr.kind else {
+        return false;
+    };
 
     // We can suggest using an inclusive range
     // (`..=`) instead only if it is the `end` that is
@@ -61,7 +67,9 @@ fn lint_overflowing_range_endpoint<'tcx>(
     };
 
     let sub_sugg = if span.lo() == lit_span.lo() {
-        let Ok(start) = cx.sess().source_map().span_to_snippet(start.span) else { return false };
+        let Ok(start) = cx.sess().source_map().span_to_snippet(start.span) else {
+            return false;
+        };
         UseInclusiveRange::WithoutParen {
             sugg: struct_expr.span.shrink_to_lo().to(lit_span.shrink_to_hi()),
             start,
@@ -316,11 +324,19 @@ fn lint_uint_literal<'tcx>(
             match par_e.kind {
                 hir::ExprKind::Cast(..) => {
                     if let ty::Char = cx.typeck_results().expr_ty(par_e).kind() {
-                        cx.emit_span_lint(
-                            OVERFLOWING_LITERALS,
-                            par_e.span,
-                            OnlyCastu8ToChar { span: par_e.span, literal: lit_val },
-                        );
+                        if lit_val <= 0xFF {
+                            cx.emit_span_lint(
+                                OVERFLOWING_LITERALS,
+                                par_e.span,
+                                OnlyCastu8ToChar { span: par_e.span, literal: lit_val },
+                            );
+                        } else {
+                            cx.emit_span_lint(
+                                OVERFLOWING_LITERALS,
+                                par_e.span,
+                                InvalidCharCast { literal: lit_val },
+                            );
+                        }
                         return;
                     }
                 }

tests/ui/cast/cast-char.rs

@@ -2,9 +2,9 @@
 
 fn main() {
     const XYZ: char = 0x1F888 as char;
-    //~^ ERROR only `u8` can be cast into `char`
+    //~^ ERROR: cannot cast
     const XY: char = 129160 as char;
-    //~^ ERROR only `u8` can be cast into `char`
+    //~^ ERROR: cannot cast
     const ZYX: char = '\u{01F888}';
     println!("{}", XYZ);
 }

tests/ui/cast/cast-char.stderr

@@ -1,20 +1,25 @@
-error: only `u8` can be cast into `char`
+error: cannot cast `129160` to `char`
   --> $DIR/cast-char.rs:4:23
    |
 LL |     const XYZ: char = 0x1F888 as char;
-   |                       ^^^^^^^^^^^^^^^ help: use a `char` literal instead: `'\u{1F888}'`
+   |                       ^^^^^^^^^^^^^^^
    |
+   = note: char must be a valid Unicode scalar value (code points in ranges [0x0, 0xD7FF] or [0xE000, 0x10FFFF]).
+   = note: The gap [0xD800, 0xDFFF] is reserved for UTF-16 surrogates and is invalid.
 note: the lint level is defined here
   --> $DIR/cast-char.rs:1:9
    |
 LL | #![deny(overflowing_literals)]
    |         ^^^^^^^^^^^^^^^^^^^^
 
-error: only `u8` can be cast into `char`
+error: cannot cast `129160` to `char`
   --> $DIR/cast-char.rs:6:22
    |
 LL |     const XY: char = 129160 as char;
-   |                      ^^^^^^^^^^^^^^ help: use a `char` literal instead: `'\u{1F888}'`
+   |                      ^^^^^^^^^^^^^^
+   |
+   = note: char must be a valid Unicode scalar value (code points in ranges [0x0, 0xD7FF] or [0xE000, 0x10FFFF]).
+   = note: The gap [0xD800, 0xDFFF] is reserved for UTF-16 surrogates and is invalid.
 
 error: aborting due to 2 previous errors