[antlr-interest] Parsing byte streams

[Johannes Luber]

Felix Schmid wrote:
> Hi,
> 
> is it possible to parse byte streams, e.g. class files, using ANTLR? I
> think I have seen the topic mentioned somewhere, maybe in the wiki, but
> can't find it now....
> 
> 
> Related to this, if it is possible, how could I skip over padding bytes
> while parsing? Say, the binary stream looks something like this:
> 
> 1 byte length of value in 4-byte words (incl. padding) || 1 byte real
> value length in bytes || value        (|| means concatenation)
> 
> so if 'value' was 'aa00aa00aa', the byte stream would look something
> like '0205aa00aa00aa000000...' and the parser had to be clever enough to
> understand that from the 8 bytes following the first byte, only the
> first 5 are the real value and the 3 bytes at the end have to be skipped
> (to hit the first byte of the next segment)....Could it be done with ANTLR?
> 
> thanks for your attention,
> 
> felix
> 

Well, the last time someone asked for this, he was told to write a
binary parser in his language of his choice. ANTLR has been written to
parse human-readable files and even if you could create a grammar for
that stuff without changing the lexer (the lexer reads probably Unicode
chars with two bytes instead one), I doubt it would be readable.

Best regards,
Johannes Luber