summaryrefslogtreecommitdiff
path: root/src
diff options
context:
space:
mode:
authordaurnimator <quae@daurnimator.com>2017-05-29 17:40:04 +1000
committerdaurnimator <quae@daurnimator.com>2017-05-29 17:54:26 +1000
commit916bbd767630181c5e7ca7cc29e0c8bdb99ccbd3 (patch)
treec92b7921e1455551cfa1cc5dfaff751a11fff0d4 /src
parentd8ead252eb4d6bded38c7a041aacfc41fa8b4263 (diff)
downloadfengari-916bbd767630181c5e7ca7cc29e0c8bdb99ccbd3.tar.gz
fengari-916bbd767630181c5e7ca7cc29e0c8bdb99ccbd3.tar.bz2
fengari-916bbd767630181c5e7ca7cc29e0c8bdb99ccbd3.zip
src/defs.js: Use .codePointAt to avoid short-comings of .charCodeAt
Diffstat (limited to 'src')
-rw-r--r--src/defs.js5
1 files changed, 2 insertions, 3 deletions
diff --git a/src/defs.js b/src/defs.js
index d1b98d0..4ea8ad7 100644
--- a/src/defs.js
+++ b/src/defs.js
@@ -197,11 +197,10 @@ const to_luastring = function(str, cache, maxBytesToWrite) {
let outIdx = 0;
let endIdx = maxBytesToWrite - 1; // -1 for string null terminator.
for (let i = 0; i < str.length; ++i) {
- // Gotcha: charCodeAt returns a 16-bit word that is a UTF-16 encoded code unit, not a Unicode code point of the character! So decode UTF16->UTF32->UTF8.
// See http://unicode.org/faq/utf_bom.html#utf16-3
// For UTF8 byte structure, see http://en.wikipedia.org/wiki/UTF-8#Description and https://www.ietf.org/rfc/rfc2279.txt and https://tools.ietf.org/html/rfc3629
- let u = str.charCodeAt(i); // possibly a lead surrogate
- if (u >= 0xD800 && u <= 0xDFFF) u = 0x10000 + ((u & 0x3FF) << 10) | (str.charCodeAt(++i) & 0x3FF);
+ let u = str.codePointAt(i);
+ if (u >= 0xD800) i++; // If it was a surrogate pair it used up two bytes
if (u <= 0x7F) {
if (outIdx >= endIdx) break;
outU8Array[outIdx++] = u;