Andreas Kling 1a46d8df5f LibTextCodec: Use String::from_utf8() when decoding UTF-8 to UTF-8 ... This way, we still perform UTF-8 validation, but don't go through the slow generic code path that rebuilds the decoded string one code point at a time. This was a bottleneck when loading a canned copy of reddit.com, which ended up being ~120 MiB large. - Time spent decoding UTF-8 before this change: 1192 ms - Time spent decoding UTF-8 after this change: 154 ms That's still a long time, but 7.7x faster is nothing to sneeze at! :^) Note that if the input fails UTF-8 validation, we still fall back to the slow path and insert replacement characters per the WHATWG Encoding spec: https://encoding.spec.whatwg.org/#utf-8-decode		2024-07-20 14:29:37 +02:00
..
CMakeLists.txt	LibTextCodec: Add GBK/GB18030 decoder	2024-05-31 07:56:26 +02:00
Decoder.cpp	LibTextCodec: Use String::from_utf8() when decoding UTF-8 to UTF-8	2024-07-20 14:29:37 +02:00
Decoder.h	LibTextCodec: Use generated lookup tables for all single byte decoders	2024-06-04 10:21:07 +02:00
indexes.json	LibTextCodec: Add GBK/GB18030 decoder	2024-05-31 07:56:26 +02:00