- From: Norm Tovey-Walsh <norm@saxonica.com>
- Date: Thu, 03 Feb 2022 19:02:05 +0000
- To: "C. M. Sperberg-McQueen" <cmsmcq@blackmesatech.com>
- Cc: public-ixml@w3.org
Received on Thursday, 3 February 2022 19:08:20 UTC
> What has changed? I wasn’t aware of the discussion from last April. I have not read all of the archives with the care I might have liked. I don’t think this is behavior that needs to be conformant in 1.0, so I’m happy to leave it closed. It seems like something that could be added later. I implemented a version of it, partly out of curiosity and partly because I have found various aspects of writing an ixml parser extremely frustrating and in the moment, this wasn’t one of them. For those interested, if I reach a point in the parse where I’ve clearly gone off the rails never to recover, I look back at the last token that was successfully parsed. If among the parses for that token was a complete sentence in the grammar, then I provide a mechanism for returning it. I also buffer up the tokens consumed from the input iterator so that a user who wishes can restart at the first bad character. If, for example, you were matching a*b* and you fed the input aaaabbbbaaaabbbb, you could extract the prefix parse “aaaabbbb” and restart the parse (or start another parse, it’s just an iterator) with “aaaabbbb”. Be seeing you, norm -- Norm Tovey-Walsh Saxonica
Received on Thursday, 3 February 2022 19:08:20 UTC