Use String::from_utf8_unchecked for encode_basestring

Author: coolreader18

vm/src/stdlib/json.rs

@@ -225,10 +225,10 @@ mod _json {
     fn encode_string(s: &str, ascii_only: bool) -> String {
         let mut buf = Vec::<u8>::with_capacity(s.len() + 2);
         super::machinery::write_json_string(s, ascii_only, &mut buf)
-            // writing to a vec can't fail
+            // SAFETY: writing to a vec can't fail
             .unwrap_or_else(|_| unsafe { std::hint::unreachable_unchecked() });
-        // TODO: verify that the implementation is correct enough to use `from_utf8_unchecked`
-        String::from_utf8(buf).expect("invalid utf-8 in json output")
+        // SAFETY: we only output valid utf8 from write_json_string
+        unsafe { String::from_utf8_unchecked(buf) }
     }
 
     #[pyfunction]

vm/src/stdlib/json/machinery.rs

@@ -28,11 +28,11 @@
 
 use std::io;
 
-static ESCAPE_CHARS: [&[u8]; 0x20] = [
-    b"\\u0000", b"\\u0001", b"\\u0002", b"\\u0003", b"\\u0004", b"\\u0005", b"\\u0006", b"\\u0007",
-    b"\\b", b"\\t", b"\\n", b"\\u000b", b"\\f", b"\\r", b"\\u000e", b"\\u000f", b"\\u0010",
-    b"\\u0011", b"\\u0012", b"\\u0013", b"\\u0014", b"\\u0015", b"\\u0016", b"\\u0017", b"\\u0018",
-    b"\\u0019", b"\\u001a", b"\\u001b", b"\\u001c", b"\\u001d", b"\\u001e", b"\\u001f",
+static ESCAPE_CHARS: [&str; 0x20] = [
+    "\\u0000", "\\u0001", "\\u0002", "\\u0003", "\\u0004", "\\u0005", "\\u0006", "\\u0007", "\\b",
+    "\\t", "\\n", "\\u000", "\\f", "\\r", "\\u000e", "\\u000f", "\\u0010", "\\u0011", "\\u0012",
+    "\\u0013", "\\u0014", "\\u0015", "\\u0016", "\\u0017", "\\u0018", "\\u0019", "\\u001a",
+    "\\u001", "\\u001c", "\\u001d", "\\u001e", "\\u001f",
 ];
 
 // This bitset represents which bytes can be copied as-is to a JSON string (0)
@@ -48,16 +48,16 @@ static NEEDS_ESCAPING_BITSET: [u64; 4] = [
 ];
 
 #[inline(always)]
-fn json_escaped_char(c: u8) -> Option<&'static [u8]> {
+fn json_escaped_char(c: u8) -> Option<&'static str> {
     let bitset_value = NEEDS_ESCAPING_BITSET[(c / 64) as usize] & (1 << (c % 64));
     if bitset_value == 0 {
         None
     } else {
         Some(match c {
             x if x < 0x20 => ESCAPE_CHARS[c as usize],
-            b'\\' => &b"\\\\"[..],
-            b'\"' => &b"\\\""[..],
-            0x7F => &b"\\u007f"[..],
+            b'\\' => "\\\\",
+            b'\"' => "\\\"",
+            0x7F => "\\u007f",
             _ => unreachable!(),
         })
     }
@@ -72,7 +72,7 @@ pub fn write_json_string<W: io::Write>(s: &str, ascii_only: bool, w: &mut W) ->
             if c.is_ascii() {
                 if let Some(escaped) = json_escaped_char(c as u8) {
                     w.write_all(&bytes[write_start_idx..idx])?;
-                    w.write_all(escaped)?;
+                    w.write_all(escaped.as_bytes())?;
                     write_start_idx = idx + 1;
                 }
             } else {
@@ -88,7 +88,7 @@ pub fn write_json_string<W: io::Write>(s: &str, ascii_only: bool, w: &mut W) ->
         for (idx, c) in s.bytes().enumerate() {
             if let Some(escaped) = json_escaped_char(c) {
                 w.write_all(&bytes[write_start_idx..idx])?;
-                w.write_all(escaped)?;
+                w.write_all(escaped.as_bytes())?;
                 write_start_idx = idx + 1;
             }
         }