From d37daeb76ad078d4a2254419cfebbc0602a50670 Mon Sep 17 00:00:00 2001
From: Justin Whear <justin.whear@gmail.com>
Date: Fri, 17 Feb 2023 14:48:10 -0800
Subject: [PATCH] Fix #2005 (#2096)

---
 src/js_printer.zig             | 49 +++++++++++++++++++++++++++++++++-
 test/bun.js/repro_2005.test.js | 12 +++++++++
 2 files changed, 60 insertions(+), 1 deletion(-)
 create mode 100644 test/bun.js/repro_2005.test.js

diff --git a/src/js_printer.zig b/src/js_printer.zig
index 48e4c62600..ce4138cd4e 100644
--- a/src/js_printer.zig
+++ b/src/js_printer.zig
@@ -2629,7 +2629,54 @@ pub fn NewPrinter(
                 p.print(" ");
             }
 
-            p.print(e.value);
+            if (comptime is_bun_platform) {
+                // Translate any non-ASCII to unicode escape sequences
+                var ascii_start: usize = 0;
+                var is_ascii = false;
+                var iter = CodepointIterator.init(e.value);
+                var cursor = CodepointIterator.Cursor{};
+                while (iter.next(&cursor)) {
+                    switch (cursor.c) {
+                        first_ascii...last_ascii => {
+                            if (!is_ascii) {
+                                ascii_start = cursor.i;
+                                is_ascii = true;
+                            }
+                        },
+                        else => {
+                            if (is_ascii) {
+                                p.print(e.value[ascii_start..cursor.i]);
+                                is_ascii = false;
+                            }
+
+                            switch (cursor.c) {
+                                0...0xFFFF => {
+                                    p.print([_]u8{
+                                        '\\',
+                                        'u',
+                                        hex_chars[cursor.c >> 12],
+                                        hex_chars[(cursor.c >> 8) & 15],
+                                        hex_chars[(cursor.c >> 4) & 15],
+                                        hex_chars[cursor.c & 15],
+                                    });
+                                },
+                                else => {
+                                    p.print("\\u{");
+                                    std.fmt.formatInt(cursor.c, 16, .lower, .{}, p) catch unreachable;
+                                    p.print("}");
+                                },
+                            }
+                        },
+                    }
+                }
+
+                if (is_ascii) {
+                    p.print(e.value[ascii_start..]);
+                }
+            } else {
+                // UTF8 sequence is fine
+                p.print(e.value);
+            }
 
             // Need a space before the next identifier to avoid it turning into flags
             p.prev_reg_exp_end = p.writer.written;
diff --git a/test/bun.js/repro_2005.test.js b/test/bun.js/repro_2005.test.js
new file mode 100644
index 0000000000..bd80ab7dd9
--- /dev/null
+++ b/test/bun.js/repro_2005.test.js
@@ -0,0 +1,12 @@
+import { it, expect } from "bun:test";
+
+it("regex literal with non-Latin1 should work", () => {
+  const text = "这是一段要替换的文字";
+
+  //Correct results: 这是一段的文字
+  expect(text.replace(new RegExp("要替换"), "")).toBe("这是一段的文字");
+
+  //Incorrect result: 这是一段要替换的文字
+  expect(text.replace(/要替换/, "")).toBe("这是一段的文字");
+
+});