From 71cdf6a920370b43e5ac0fd8d1da8aba22cd2044 Mon Sep 17 00:00:00 2001
From: Nathan Shively-Sanders <293473+sandersn@users.noreply.github.com>
Date: Thu, 20 May 2021 14:39:31 -0700
Subject: [PATCH] Private identifiers use standard identifier scanning (#44184)

* Private identifiers use standard identifer scanning

Previously they used an old copy of the identifier scanning code that
didn't handle extended unicode yet.

* gotta fix that const lint
---
 src/compiler/scanner.ts                       | 15 +----
 .../extendedUnicodePlaneIdentifiers.js        | 13 +++++
 .../extendedUnicodePlaneIdentifiers.symbols   | 56 +++++++++++--------
 .../extendedUnicodePlaneIdentifiers.types     | 12 ++++
 .../extendedUnicodePlaneIdentifiers.ts        |  5 ++
 5 files changed, 65 insertions(+), 36 deletions(-)

diff --git a/src/compiler/scanner.ts b/src/compiler/scanner.ts
index ba449e81b96..7ce7f97cf7d 100644
--- a/src/compiler/scanner.ts
+++ b/src/compiler/scanner.ts
@@ -1610,7 +1610,7 @@ namespace ts {
                 if (pos >= end) {
                     return token = SyntaxKind.EndOfFileToken;
                 }
-                let ch = codePointAt(text, pos);
+                const ch = codePointAt(text, pos);
 
                 // Special handling for shebang
                 if (ch === CharacterCodes.hash && pos === 0 && isShebangTrivia(text, pos)) {
@@ -2044,18 +2044,7 @@ namespace ts {
                             return token = SyntaxKind.Unknown;
                         }
                         pos++;
-                        if (isIdentifierStart(ch = text.charCodeAt(pos), languageVersion)) {
-                            pos++;
-                            while (pos < end && isIdentifierPart(ch = text.charCodeAt(pos), languageVersion)) pos++;
-                            tokenValue = text.substring(tokenPos, pos);
-                            if (ch === CharacterCodes.backslash) {
-                                tokenValue += scanIdentifierParts();
-                            }
-                        }
-                        else {
-                            tokenValue = "#";
-                            error(Diagnostics.Invalid_character);
-                        }
+                        scanIdentifier(codePointAt(text, pos), languageVersion);
                         return token = SyntaxKind.PrivateIdentifier;
                     default:
                         const identifierKind = scanIdentifier(ch, languageVersion);
diff --git a/tests/baselines/reference/extendedUnicodePlaneIdentifiers.js b/tests/baselines/reference/extendedUnicodePlaneIdentifiers.js
index 7a5a5d9cc04..36225682cf2 100644
--- a/tests/baselines/reference/extendedUnicodePlaneIdentifiers.js
+++ b/tests/baselines/reference/extendedUnicodePlaneIdentifiers.js
@@ -3,6 +3,11 @@ const 𝑚 = 4;
 const 𝑀 = 5;
 console.log(𝑀 + 𝑚); // 9
 
+class K {
+    #𝑚 = 4;
+    #𝑀 = 5;
+}
+
 // lower 8 bits look like 'a'
 const ၡ = 6;
 console.log(ၡ ** ၡ);
@@ -39,9 +44,17 @@ const 𝓮𡚭𓀺ⱱ = "ok";
 
 
 //// [extendedUnicodePlaneIdentifiers.js]
+var _K_𝑚, _K_𝑀;
 const 𝑚 = 4;
 const 𝑀 = 5;
 console.log(𝑀 + 𝑚); // 9
+class K {
+    constructor() {
+        _K_𝑚.set(this, 4);
+        _K_𝑀.set(this, 5);
+    }
+}
+_K_𝑚 = new WeakMap(), _K_𝑀 = new WeakMap();
 // lower 8 bits look like 'a'
 const ၡ = 6;
 console.log(ၡ ** ၡ);
diff --git a/tests/baselines/reference/extendedUnicodePlaneIdentifiers.symbols b/tests/baselines/reference/extendedUnicodePlaneIdentifiers.symbols
index 170b82da7e9..c5fff26c53f 100644
--- a/tests/baselines/reference/extendedUnicodePlaneIdentifiers.symbols
+++ b/tests/baselines/reference/extendedUnicodePlaneIdentifiers.symbols
@@ -12,80 +12,90 @@ console.log(𝑀 + 𝑚); // 9
 >𝑀 : Symbol(𝑀, Decl(extendedUnicodePlaneIdentifiers.ts, 1, 5))
 >𝑚 : Symbol(𝑚, Decl(extendedUnicodePlaneIdentifiers.ts, 0, 5))
 
+class K {
+>K : Symbol(K, Decl(extendedUnicodePlaneIdentifiers.ts, 2, 21))
+
+    #𝑚 = 4;
+>#𝑚 : Symbol(K[#𝑚], Decl(extendedUnicodePlaneIdentifiers.ts, 4, 9))
+
+    #𝑀 = 5;
+>#𝑀 : Symbol(K[#𝑀], Decl(extendedUnicodePlaneIdentifiers.ts, 5, 12))
+}
+
 // lower 8 bits look like 'a'
 const ၡ = 6;
->ၡ : Symbol(ၡ, Decl(extendedUnicodePlaneIdentifiers.ts, 5, 5))
+>ၡ : Symbol(ၡ, Decl(extendedUnicodePlaneIdentifiers.ts, 10, 5))
 
 console.log(ၡ ** ၡ);
 >console.log : Symbol(Console.log, Decl(lib.dom.d.ts, --, --))
 >console : Symbol(console, Decl(lib.dom.d.ts, --, --))
 >log : Symbol(Console.log, Decl(lib.dom.d.ts, --, --))
->ၡ : Symbol(ၡ, Decl(extendedUnicodePlaneIdentifiers.ts, 5, 5))
->ၡ : Symbol(ၡ, Decl(extendedUnicodePlaneIdentifiers.ts, 5, 5))
+>ၡ : Symbol(ၡ, Decl(extendedUnicodePlaneIdentifiers.ts, 10, 5))
+>ၡ : Symbol(ၡ, Decl(extendedUnicodePlaneIdentifiers.ts, 10, 5))
 
 // lower 8 bits aren't a valid unicode character
 const ဒ = 7;
->ဒ : Symbol(ဒ, Decl(extendedUnicodePlaneIdentifiers.ts, 9, 5))
+>ဒ : Symbol(ဒ, Decl(extendedUnicodePlaneIdentifiers.ts, 14, 5))
 
 console.log(ဒ ** ဒ);
 >console.log : Symbol(Console.log, Decl(lib.dom.d.ts, --, --))
 >console : Symbol(console, Decl(lib.dom.d.ts, --, --))
 >log : Symbol(Console.log, Decl(lib.dom.d.ts, --, --))
->ဒ : Symbol(ဒ, Decl(extendedUnicodePlaneIdentifiers.ts, 9, 5))
->ဒ : Symbol(ဒ, Decl(extendedUnicodePlaneIdentifiers.ts, 9, 5))
+>ဒ : Symbol(ဒ, Decl(extendedUnicodePlaneIdentifiers.ts, 14, 5))
+>ဒ : Symbol(ဒ, Decl(extendedUnicodePlaneIdentifiers.ts, 14, 5))
 
 // a mix, for good measure
 const ဒၡ𝑀 = 7;
->ဒၡ𝑀 : Symbol(ဒၡ𝑀, Decl(extendedUnicodePlaneIdentifiers.ts, 13, 5))
+>ဒၡ𝑀 : Symbol(ဒၡ𝑀, Decl(extendedUnicodePlaneIdentifiers.ts, 18, 5))
 
 console.log(ဒၡ𝑀 ** ဒၡ𝑀);
 >console.log : Symbol(Console.log, Decl(lib.dom.d.ts, --, --))
 >console : Symbol(console, Decl(lib.dom.d.ts, --, --))
 >log : Symbol(Console.log, Decl(lib.dom.d.ts, --, --))
->ဒၡ𝑀 : Symbol(ဒၡ𝑀, Decl(extendedUnicodePlaneIdentifiers.ts, 13, 5))
->ဒၡ𝑀 : Symbol(ဒၡ𝑀, Decl(extendedUnicodePlaneIdentifiers.ts, 13, 5))
+>ဒၡ𝑀 : Symbol(ဒၡ𝑀, Decl(extendedUnicodePlaneIdentifiers.ts, 18, 5))
+>ဒၡ𝑀 : Symbol(ဒၡ𝑀, Decl(extendedUnicodePlaneIdentifiers.ts, 18, 5))
 
 const ၡ𝑀ဒ = 7;
->ၡ𝑀ဒ : Symbol(ၡ𝑀ဒ, Decl(extendedUnicodePlaneIdentifiers.ts, 16, 5))
+>ၡ𝑀ဒ : Symbol(ၡ𝑀ဒ, Decl(extendedUnicodePlaneIdentifiers.ts, 21, 5))
 
 console.log(ၡ𝑀ဒ ** ၡ𝑀ဒ);
 >console.log : Symbol(Console.log, Decl(lib.dom.d.ts, --, --))
 >console : Symbol(console, Decl(lib.dom.d.ts, --, --))
 >log : Symbol(Console.log, Decl(lib.dom.d.ts, --, --))
->ၡ𝑀ဒ : Symbol(ၡ𝑀ဒ, Decl(extendedUnicodePlaneIdentifiers.ts, 16, 5))
->ၡ𝑀ဒ : Symbol(ၡ𝑀ဒ, Decl(extendedUnicodePlaneIdentifiers.ts, 16, 5))
+>ၡ𝑀ဒ : Symbol(ၡ𝑀ဒ, Decl(extendedUnicodePlaneIdentifiers.ts, 21, 5))
+>ၡ𝑀ဒ : Symbol(ၡ𝑀ဒ, Decl(extendedUnicodePlaneIdentifiers.ts, 21, 5))
 
 const 𝑀ဒၡ = 7;
->𝑀ဒၡ : Symbol(𝑀ဒၡ, Decl(extendedUnicodePlaneIdentifiers.ts, 19, 5))
+>𝑀ဒၡ : Symbol(𝑀ဒၡ, Decl(extendedUnicodePlaneIdentifiers.ts, 24, 5))
 
 console.log(𝑀ဒၡ ** 𝑀ဒၡ);
 >console.log : Symbol(Console.log, Decl(lib.dom.d.ts, --, --))
 >console : Symbol(console, Decl(lib.dom.d.ts, --, --))
 >log : Symbol(Console.log, Decl(lib.dom.d.ts, --, --))
->𝑀ဒၡ : Symbol(𝑀ဒၡ, Decl(extendedUnicodePlaneIdentifiers.ts, 19, 5))
->𝑀ဒၡ : Symbol(𝑀ဒၡ, Decl(extendedUnicodePlaneIdentifiers.ts, 19, 5))
+>𝑀ဒၡ : Symbol(𝑀ဒၡ, Decl(extendedUnicodePlaneIdentifiers.ts, 24, 5))
+>𝑀ဒၡ : Symbol(𝑀ဒၡ, Decl(extendedUnicodePlaneIdentifiers.ts, 24, 5))
 
 const 𝓱𝓮𝓵𝓵𝓸 = "𝔀𝓸𝓻𝓵𝓭";
->𝓱𝓮𝓵𝓵𝓸 : Symbol(𝓱𝓮𝓵𝓵𝓸, Decl(extendedUnicodePlaneIdentifiers.ts, 22, 5))
+>𝓱𝓮𝓵𝓵𝓸 : Symbol(𝓱𝓮𝓵𝓵𝓸, Decl(extendedUnicodePlaneIdentifiers.ts, 27, 5))
 
 const Ɐⱱ = "ok"; // BMP
->Ɐⱱ : Symbol(Ɐⱱ, Decl(extendedUnicodePlaneIdentifiers.ts, 24, 5))
+>Ɐⱱ : Symbol(Ɐⱱ, Decl(extendedUnicodePlaneIdentifiers.ts, 29, 5))
 
 const 𓀸𓀹𓀺 = "ok"; // SMP
->𓀸𓀹𓀺 : Symbol(𓀸𓀹𓀺, Decl(extendedUnicodePlaneIdentifiers.ts, 26, 5))
+>𓀸𓀹𓀺 : Symbol(𓀸𓀹𓀺, Decl(extendedUnicodePlaneIdentifiers.ts, 31, 5))
 
 const 𡚭𡚮𡚯 = "ok"; // SIP
->𡚭𡚮𡚯 : Symbol(𡚭𡚮𡚯, Decl(extendedUnicodePlaneIdentifiers.ts, 28, 5))
+>𡚭𡚮𡚯 : Symbol(𡚭𡚮𡚯, Decl(extendedUnicodePlaneIdentifiers.ts, 33, 5))
 
 const 𡚭𓀺ⱱ𝓮 = "ok";
->𡚭𓀺ⱱ𝓮 : Symbol(𡚭𓀺ⱱ𝓮, Decl(extendedUnicodePlaneIdentifiers.ts, 30, 5))
+>𡚭𓀺ⱱ𝓮 : Symbol(𡚭𓀺ⱱ𝓮, Decl(extendedUnicodePlaneIdentifiers.ts, 35, 5))
 
 const 𓀺ⱱ𝓮𡚭 = "ok";
->𓀺ⱱ𝓮𡚭 : Symbol(𓀺ⱱ𝓮𡚭, Decl(extendedUnicodePlaneIdentifiers.ts, 32, 5))
+>𓀺ⱱ𝓮𡚭 : Symbol(𓀺ⱱ𝓮𡚭, Decl(extendedUnicodePlaneIdentifiers.ts, 37, 5))
 
 const ⱱ𝓮𡚭𓀺 = "ok";
->ⱱ𝓮𡚭𓀺 : Symbol(ⱱ𝓮𡚭𓀺, Decl(extendedUnicodePlaneIdentifiers.ts, 34, 5))
+>ⱱ𝓮𡚭𓀺 : Symbol(ⱱ𝓮𡚭𓀺, Decl(extendedUnicodePlaneIdentifiers.ts, 39, 5))
 
 const 𝓮𡚭𓀺ⱱ = "ok";
->𝓮𡚭𓀺ⱱ : Symbol(𝓮𡚭𓀺ⱱ, Decl(extendedUnicodePlaneIdentifiers.ts, 36, 5))
+>𝓮𡚭𓀺ⱱ : Symbol(𝓮𡚭𓀺ⱱ, Decl(extendedUnicodePlaneIdentifiers.ts, 41, 5))
 
diff --git a/tests/baselines/reference/extendedUnicodePlaneIdentifiers.types b/tests/baselines/reference/extendedUnicodePlaneIdentifiers.types
index 8c96a80de7c..f9b4605d4f4 100644
--- a/tests/baselines/reference/extendedUnicodePlaneIdentifiers.types
+++ b/tests/baselines/reference/extendedUnicodePlaneIdentifiers.types
@@ -16,6 +16,18 @@ console.log(𝑀 + 𝑚); // 9
 >𝑀 : 5
 >𝑚 : 4
 
+class K {
+>K : K
+
+    #𝑚 = 4;
+>#𝑚 : number
+>4 : 4
+
+    #𝑀 = 5;
+>#𝑀 : number
+>5 : 5
+}
+
 // lower 8 bits look like 'a'
 const ၡ = 6;
 >ၡ : 6
diff --git a/tests/cases/compiler/extendedUnicodePlaneIdentifiers.ts b/tests/cases/compiler/extendedUnicodePlaneIdentifiers.ts
index 9210ee3b7c5..59baad082b5 100644
--- a/tests/cases/compiler/extendedUnicodePlaneIdentifiers.ts
+++ b/tests/cases/compiler/extendedUnicodePlaneIdentifiers.ts
@@ -3,6 +3,11 @@ const 𝑚 = 4;
 const 𝑀 = 5;
 console.log(𝑀 + 𝑚); // 9
 
+class K {
+    #𝑚 = 4;
+    #𝑀 = 5;
+}
+
 // lower 8 bits look like 'a'
 const ၡ = 6;
 console.log(ၡ ** ၡ);