gh-95865: Further reduce quote_from_bytes memory consumption

[gpshead]

Further reduce quote_from_bytes memory consumption on large input values. Based on Dennis Sweeney's chunking idea.

• Issue: quote_from_bytes uses a lot of memory for larger bytestrings #95865

[gpshead]

On large values, this consumes ~half the memory.

[sweeneyde]

This looks like a correct, comprehensible change, my only wonder would be whether some sort of regex-based approach would be more efficient for this quote direction as well, i.e., have a regex recognize strings of safe characters.

[sweeneyde]

It may also depend on what sorts of strings are typically quoted/unquoted. Different texts/corpuses probably have different performance characteristics, and I am not all that familiar with how this gets used in the wild.

[gpshead]

there are likely further optimization opportunities. but as this does use notably less memory already, lets go ahead and do it. I don't like introducing regexes when possible that could be tested and considered as a separate change if anyone ever cares here.