From adbe64fffc56d5a63fb92bfde3d863d7d90b68ee Mon Sep 17 00:00:00 2001
From: Kampfkarren <boynedmaster@gmail.com>
Date: Mon, 22 Aug 2022 15:22:28 -0700
Subject: [PATCH] Support unicode escapes

---
 Ast/src/Lexer.cpp                  |  9 +++++++++
 Compiler/src/Compiler.cpp          |  2 +-
 tests/Lexer.test.cpp               | 13 +++++++++++++
 tests/conformance/stringinterp.lua |  2 ++
 4 files changed, 25 insertions(+), 1 deletion(-)

diff --git a/Ast/src/Lexer.cpp b/Ast/src/Lexer.cpp
index 70612ca1..870e11c2 100644
--- a/Ast/src/Lexer.cpp
+++ b/Ast/src/Lexer.cpp
@@ -624,6 +624,15 @@ Lexeme Lexer::readInterpolatedStringSection(Position start, Lexeme::Type formatT
             return Lexeme(Location(start, position()), Lexeme::BrokenString);
 
         case '\\':
+            // Allow for \u{}, which would otherwise be consumed by looking for {
+            if (peekch(1) == 'u' && peekch(2) == '{')
+            {
+                consume(); // backslash
+                consume(); // u
+                consume(); // {
+                break;
+            }
+
             readBackslashInString();
             break;
 
diff --git a/Compiler/src/Compiler.cpp b/Compiler/src/Compiler.cpp
index 17256e15..666cca54 100644
--- a/Compiler/src/Compiler.cpp
+++ b/Compiler/src/Compiler.cpp
@@ -1639,7 +1639,7 @@ struct Compiler
 
         RegScope rs(this);
 
-        uint8_t baseReg = allocReg(expr, 2 + expr->expressions.size);
+        uint8_t baseReg = allocReg(expr, uint8_t(2 + expr->expressions.size));
 
         emitLoadK(baseReg, formatStringIndex);
 
diff --git a/tests/Lexer.test.cpp b/tests/Lexer.test.cpp
index 7330aba6..9ff06bc9 100644
--- a/tests/Lexer.test.cpp
+++ b/tests/Lexer.test.cpp
@@ -175,4 +175,17 @@ TEST_CASE("string_interpolation_unmatched_brace")
     CHECK_EQ(lexer.next().type, '}');
 }
 
+TEST_CASE("string_interpolation_with_unicode_escape")
+{
+    ScopedFastFlag sff{"LuauInterpolatedStringBaseSupport", true};
+
+    const std::string testInput = R"(`\u{1F41B}`)";
+    Luau::Allocator alloc;
+    AstNameTable table(alloc);
+    Lexer lexer(testInput.c_str(), testInput.size(), table);
+
+    CHECK_EQ(lexer.next().type, Lexeme::InterpStringSimple);
+    CHECK_EQ(lexer.next().type, Lexeme::Eof);
+}
+
 TEST_SUITE_END();
diff --git a/tests/conformance/stringinterp.lua b/tests/conformance/stringinterp.lua
index 9bb29a7e..539505f4 100644
--- a/tests/conformance/stringinterp.lua
+++ b/tests/conformance/stringinterp.lua
@@ -61,4 +61,6 @@ end
 assertEq(identity`text`, "text")
 assertEq(identity`foo{"bar"}`, "foobar")
 
+assertEq(`\u{0041}\t`, "A\t")
+
 return "OK"