Decode multibyte escapes immediately

We forgot to decode (i.e. turn into nice wchar_t codepoints)
"byte_literal" escape sequences. This meant that e.g.

```fish
string match ö \Xc3\Xb6

math 5 \X2b 5
```

didn't work, but `math 5 \x2b 5` did, and would print the wonderful
error:

```
math: Error: Missing operator
'5 + 5'
   ^
```

So, instead, we decode eagerly.

src/common.cpp

@@ -1149,6 +1149,10 @@ maybe_t<size_t> read_unquoted_escape(const wchar_t *input, wcstring *result, boo
     bool errored = false;
     size_t in_pos = 1;  // in_pos always tracks the next character to read (and therefore the number
                         // of characters read so far)
+
+    // For multibyte \X sequences.
+    std::string byte_buff;
+    while (!errored) {
         const wchar_t c = input[in_pos++];
         switch (c) {
                 // A null character after a backslash is an error.
@@ -1207,8 +1211,8 @@ maybe_t<size_t> read_unquoted_escape(const wchar_t *input, wcstring *result, boo
                     default: {
                         base = 8;
                         chars = 3;
-                    // Note that in_pos currently is just after the first post-backslash character;
-                    // we want to start our escape from there.
+                        // Note that in_pos currently is just after the first post-backslash
+                        // character; we want to start our escape from there.
                         assert(in_pos > 0);
                         in_pos--;
                         break;
@@ -1228,8 +1232,20 @@ maybe_t<size_t> read_unquoted_escape(const wchar_t *input, wcstring *result, boo
                 }
 
                 if (!errored && res <= max_val) {
-                result_char_or_none =
-                    static_cast<wchar_t>((byte_literal ? ENCODE_DIRECT_BASE : 0) + res);
+                    if (byte_literal) {
+                        // Multibyte encodings necessitate that we keep adjacent byte escapes.
+                        // - `\Xc3\Xb6` is "ö", but only together.
+                        // (this assumes a valid codepoint can't consist of multiple bytes
+                        // that are valid on their own, which is true for UTF-8)
+                        byte_buff.push_back(static_cast<char>(res));
+                        result_char_or_none = none();
+                        if (input[in_pos] == L'\\' && input[in_pos + 1] == L'X') {
+                            in_pos++;
+                            continue;
+                        }
+                    } else {
+                        result_char_or_none = static_cast<wchar_t>(res);
+                    }
                 } else {
                     errored = true;
                 }
@@ -1300,10 +1316,18 @@ maybe_t<size_t> read_unquoted_escape(const wchar_t *input, wcstring *result, boo
             }
         }
 
-    if (!errored && result_char_or_none.has_value()) {
+        if (errored) return none();
+
+        if (!byte_buff.empty()) {
+            result->append(str2wcstring(byte_buff));
+        }
+
+        break;
+    }
+
+    if (result_char_or_none.has_value()) {
         result->push_back(*result_char_or_none);
     }
-    if (errored) return none();
 
     return in_pos;
 }

tests/checks/locale.fish

@@ -88,3 +88,12 @@ env LC_ALL=C $fish -c 'echo -n Y\u00FCY' | display_bytes
 env LC_ALL=C $fish -c 'echo -n T\u01FDT' | display_bytes
 #CHECK: 0000000 124 077 124
 #CHECK: 0000003
+
+string match ö \Xc3\Xb6
+#CHECK: ö
+
+math 5 \X2b 5
+#CHECK: 10
+
+math 7 \x2b 7
+#CHECK: 14