Parse multiple versions of a document format?

[Rui Maciel]

I'm writing a parser for a document format which has a couple of version out. 
The document format basically consists of a header field, where the document 
format version is specified and whose format is the same across all format 
versions, followed by multiple fields which differ between versions.

Having that case in mind, what is the best strategy to write a bison parser 
that is capable of parsing multiple versions of a document format? It would be 
possible to write a single bison parser that support all document version, 
which has the potential of ending up an ungodly, unmanageable pile of grammar 
definitions. It would also be possible to rely on a parser to process only the 
header field and, when finishing parsing it, check the version number and start 
parsing the document from that point on with a different parser which is 
specific to the version number. Yet, this seems to be a bit of a hack.

So, according to your experience, what's the best strategy to parse multiple 
versions of a document format?


Thanks in advance,
Rui Maciel