Improve diagnostics if a character literal contains combining marks

3 files changed, 90 insertions, 0 deletions

diff --git a/src/test/ui/parser/unicode-character-literal.fixed b/src/test/ui/parser/unicode-character-literal.fixed
new file mode 100644
index 00000000000..26ef5ffa11a
--- /dev/null
+++ b/src/test/ui/parser/unicode-character-literal.fixed
@@ -0,0 +1,21 @@
+// Regression test for #88684: Improve diagnostics for combining marks
+// in character literals.
+
+// run-rustfix
+
+fn main() {
+    let _spade = "♠️";
+    //~^ ERROR: character literal may only contain one codepoint
+    //~| NOTE: this `♠` is followed by the combining mark `\u{fe0f}`
+    //~| HELP: if you meant to write a `str` literal, use double quotes
+
+    let _s = "ṩ̂̊";
+    //~^ ERROR: character literal may only contain one codepoint
+    //~| NOTE: this `s` is followed by the combining marks `\u{323}\u{307}\u{302}\u{30a}`
+    //~| HELP: if you meant to write a `str` literal, use double quotes
+
+    let _a = 'Å';
+    //~^ ERROR: character literal may only contain one codepoint
+    //~| NOTE: this `A` is followed by the combining mark `\u{30a}`
+    //~| HELP: consider using the normalized form `\u{c5}` of this character
+}
diff --git a/src/test/ui/parser/unicode-character-literal.rs b/src/test/ui/parser/unicode-character-literal.rs
new file mode 100644
index 00000000000..d331522c04c
--- /dev/null
+++ b/src/test/ui/parser/unicode-character-literal.rs
@@ -0,0 +1,21 @@
+// Regression test for #88684: Improve diagnostics for combining marks
+// in character literals.
+
+// run-rustfix
+
+fn main() {
+    let _spade = '♠️';
+    //~^ ERROR: character literal may only contain one codepoint
+    //~| NOTE: this `♠` is followed by the combining mark `\u{fe0f}`
+    //~| HELP: if you meant to write a `str` literal, use double quotes
+
+    let _s = 'ṩ̂̊';
+    //~^ ERROR: character literal may only contain one codepoint
+    //~| NOTE: this `s` is followed by the combining marks `\u{323}\u{307}\u{302}\u{30a}`
+    //~| HELP: if you meant to write a `str` literal, use double quotes
+
+    let _a = 'Å';
+    //~^ ERROR: character literal may only contain one codepoint
+    //~| NOTE: this `A` is followed by the combining mark `\u{30a}`
+    //~| HELP: consider using the normalized form `\u{c5}` of this character
+}
diff --git a/src/test/ui/parser/unicode-character-literal.stderr b/src/test/ui/parser/unicode-character-literal.stderr
new file mode 100644
index 00000000000..5cd3bd0fe69
--- /dev/null
+++ b/src/test/ui/parser/unicode-character-literal.stderr
@@ -0,0 +1,48 @@
+error: character literal may only contain one codepoint
+  --> $DIR/unicode-character-literal.rs:7:18
+   |
+LL |     let _spade = '♠️';
+   |                  ^^^
+   |
+note: this `♠` is followed by the combining mark `\u{fe0f}`
+  --> $DIR/unicode-character-literal.rs:7:19
+   |
+LL |     let _spade = '♠️';
+   |                   ^
+help: if you meant to write a `str` literal, use double quotes
+   |
+LL |     let _spade = "♠️";
+   |                  ~~~
+
+error: character literal may only contain one codepoint
+  --> $DIR/unicode-character-literal.rs:12:14
+   |
+LL |     let _s = 'ṩ̂̊';
+   |              ^^^
+   |
+note: this `s` is followed by the combining marks `\u{323}\u{307}\u{302}\u{30a}`
+  --> $DIR/unicode-character-literal.rs:12:15
+   |
+LL |     let _s = 'ṩ̂̊';
+   |               ^
+help: if you meant to write a `str` literal, use double quotes
+   |
+LL |     let _s = "ṩ̂̊";
+   |              ~~~
+
+error: character literal may only contain one codepoint
+  --> $DIR/unicode-character-literal.rs:17:14
+   |
+LL |     let _a = 'Å';
+   |              ^-^
+   |               |
+   |               help: consider using the normalized form `\u{c5}` of this character: `Å`
+   |
+note: this `A` is followed by the combining mark `\u{30a}`
+  --> $DIR/unicode-character-literal.rs:17:15
+   |
+LL |     let _a = 'Å';
+   |               ^
+
+error: aborting due to 3 previous errors
+