Improve WebM detection #486
Conversation
| @@ -736,7 +736,8 @@ async function _fromTokenizer(tokenizer) { | |||
| while (children > 0) { | |||
| const element = await readElement(); | |||
| if (element.id === 0x42_82) { | |||
| return tokenizer.readToken(new Token.StringType(element.len, 'utf-8')); // Return DocType | |||
| const rawValue = await tokenizer.readToken(new Token.StringType(element.len, 'utf-8')); | |||
| return rawValue.replace(/\00.*$/g, ''); // Return DocType | |||
There was a problem hiding this comment.
Does it document the maximum amount of null characters there could be? Would be nice to have a limit in place so it wouldn't hang on faulty files that has too many null characters.
There was a problem hiding this comment.
There is no maximum, it's used as a kind of padding.
The maximum length is of the string read is already terminated by element.len.
There was a problem hiding this comment.
Ah you mean check element.len? The length can exceed the JavaScript number length and is encoded a specific way (VINT examples).
At that point the assumption is already it is EBML and starts to consume the tokenizer
and iterate through the EBML elements.
You could say, the docType must be a relative short value, but then we already matched 0x1A, 0x45, 0xDF, 0xA3 & 0x42_82. Extremely unlikely we hit that point without the format being EBML.
Fixes recognition of WebM format. Resolves: #485
Borewit
changed the title
sindresorhus
changed the title
Fixes recognition of WebM format.
Fixes: #485