Add NDRON and RON text-sequence support

[delaneyj]

Add the two RON v1 streaming profiles after #1 lands.

NDRON

The NDJSON counterpart:

RON-text LF

• UTF-8, one single-line RON value per record.
• Encoder emits LF; parser accepts LF and CRLF.
• Reject embedded raw CR and an unterminated final record.
• Empty-line handling is documented and configurable.

Recommended media type: application/x-ndron; extension: .ndron.

RON text sequences

The RFC 7464 counterpart:

RS RON-text LF

• Incremental, UTF-8 RON with binary RS framing.
• Continue at the next RS after malformed or truncated elements.
• Report invalid preamble bytes; ignore consecutive RS markers.
• LF-less wrapped/quoted self-delimited values may be accepted.
• Drop LF-less bare scalars and root-elided objects as potentially truncated.
• Never publish partial values from a failed element.

Recommended media type: application/ron-seq.

The escape correction in #1 makes framing safe: LF, CR, and RS string values encode as \n, \r, and \u001e.

Acceptance:

• Expose idiomatic incremental reader/writer APIs for both profiles.
• Enforce configurable record-size and nesting limits.
• Consume testdata/sequences/manifest.json after updating the corpus submodule.
• Exact-match valid stream bytes and pass all malformed/truncation recovery cases.

Reference: https://www.rfc-editor.org/rfc/rfc7464