Rollup merge of rust-lang#67810 - crlf0710:uncommon_codepoints_lint, …

…r=Manishearth

Implement uncommon_codepoints lint.

Part of rust-lang#55467 .  The checks of `$crate` and `{{root}}` are very unfortunate. But i'm not sure where they belongs to.

Cargo.lock

@@ -3642,6 +3642,7 @@ dependencies = [
  "rustc_span",
  "rustc_target",
  "syntax",
+ "unicode-security",
 ]
 
 [[package]]
@@ -4940,6 +4941,21 @@ dependencies = [
  "smallvec 1.0.0",
 ]
 
+[[package]]
+name = "unicode-script"
+version = "0.4.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5b2c5c29e805da6817f5af6a627d65adb045cebf05cccd5a3493d6109454391c"
+
+[[package]]
+name = "unicode-security"
+version = "0.0.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c49d35967fa037b881acc34ef717c38c4b5560eba10e3685271b3f530bb19634"
+dependencies = [
+ "unicode-script",
+]
+
 [[package]]
 name = "unicode-segmentation"
 version = "1.6.0"

src/librustc_lint/Cargo.toml

@@ -10,6 +10,7 @@ path = "lib.rs"
 
 [dependencies]
 log = "0.4"
+unicode-security = "0.0.2"
 rustc = { path = "../librustc" }
 rustc_target = { path = "../librustc_target" }
 syntax = { path = "../libsyntax" }

src/librustc_lint/non_ascii_idents.rs

@@ -7,15 +7,32 @@ declare_lint! {
     "detects non-ASCII identifiers"
 }
 
-declare_lint_pass!(NonAsciiIdents => [NON_ASCII_IDENTS]);
+declare_lint! {
+    pub UNCOMMON_CODEPOINTS,
+    Warn,
+    "detects uncommon Unicode codepoints in identifiers"
+}
+
+declare_lint_pass!(NonAsciiIdents => [NON_ASCII_IDENTS, UNCOMMON_CODEPOINTS]);
 
 impl EarlyLintPass for NonAsciiIdents {
     fn check_ident(&mut self, cx: &EarlyContext<'_>, ident: ast::Ident) {
-        if !ident.name.as_str().is_ascii() {
+        use unicode_security::GeneralSecurityProfile;
+        let name_str = ident.name.as_str();
+        if name_str.is_ascii() {
+            return;
+        }
+        cx.struct_span_lint(
+            NON_ASCII_IDENTS,
+            ident.span,
+            "identifier contains non-ASCII characters",
+        )
+        .emit();
+        if !name_str.chars().all(GeneralSecurityProfile::identifier_allowed) {
             cx.struct_span_lint(
-                NON_ASCII_IDENTS,
+                UNCOMMON_CODEPOINTS,
                 ident.span,
-                "identifier contains non-ASCII characters",
+                "identifier contains uncommon Unicode codepoints",
             )
             .emit();
         }

src/test/ui/lint/rfc-2457-non-ascii-idents/lint-uncommon-codepoints.rs

@@ -0,0 +1,11 @@
+#![feature(non_ascii_idents)]
+#![deny(uncommon_codepoints)]
+
+const µ: f64 = 0.000001; //~ ERROR identifier contains uncommon Unicode codepoints
+
+fn dijkstra() {} //~ ERROR identifier contains uncommon Unicode codepoints
+
+fn main() {
+    let ㇻㇲㇳ = "rust"; //~ ERROR identifier contains uncommon Unicode codepoints
+    println!("{}", ㇻㇲㇳ); //~ ERROR identifier contains uncommon Unicode codepoints
+}

src/test/ui/lint/rfc-2457-non-ascii-idents/lint-uncommon-codepoints.stderr

@@ -0,0 +1,32 @@
+error: identifier contains uncommon Unicode codepoints
+  --> $DIR/lint-uncommon-codepoints.rs:4:7
+   |
+LL | const µ: f64 = 0.000001;
+   |       ^
+   |
+note: lint level defined here
+  --> $DIR/lint-uncommon-codepoints.rs:2:9
+   |
+LL | #![deny(uncommon_codepoints)]
+   |         ^^^^^^^^^^^^^^^^^^^
+
+error: identifier contains uncommon Unicode codepoints
+  --> $DIR/lint-uncommon-codepoints.rs:6:4
+   |
+LL | fn dijkstra() {}
+   |    ^^^^^^^
+
+error: identifier contains uncommon Unicode codepoints
+  --> $DIR/lint-uncommon-codepoints.rs:9:9
+   |
+LL |     let ㇻㇲㇳ = "rust";
+   |         ^^^^^^
+
+error: identifier contains uncommon Unicode codepoints
+  --> $DIR/lint-uncommon-codepoints.rs:10:20
+   |
+LL |     println!("{}", ㇻㇲㇳ);
+   |                    ^^^^^^
+
+error: aborting due to 4 previous errors
+

src/tools/tidy/src/deps.rs

@@ -171,6 +171,8 @@ const WHITELIST: &[Crate<'_>] = &[
     Crate("thread_local"),
     Crate("ucd-util"),
     Crate("unicode-normalization"),
+    Crate("unicode-script"),
+    Crate("unicode-security"),
     Crate("unicode-width"),
     Crate("unicode-xid"),
     Crate("unreachable"),