Re: New JSON Parser

On Mon, Jul 27, 2020 at 1:32 PM Nathan Andelin <nandelin@xxxxxxxxx> wrote:

My inclination is to always return a meaningful reference regardless of
unexpected characters found in the stream, so that any valid content can be
extracted or reported. That's my idea of error forgiveness.

That's a principle that you're trying to operate under, which seems
fine, but it doesn't say much about how you determine what content is
"valid" in the face of malformed input.

One thing that I think most of us could imagine is that everything is
considered valid until you encounter an unexpected character, and once
that happens, the rest of the input is ignored. If you haven't stopped
at a "convenient" place, then you discard whatever it is you were in
the middle of parsing until you do get back to a convenient stopping
place, and from there, you effectively close whatever open brackets
you have outstanding.

In the case of Jon's example:

[ "name" : X , "another" : 12 ]

The colon character encountered in position 9 would not be expected.

Right, and so it's weird that he wound up with a node with a value of
X, because X is past the unexpected character. How/why did the parser
pick up the X, and why would it be considered the next element after
"name"?

John Y.