Improve parser regex performance

[colin-grant-work]

If a large message comes in many parts, the current code of the parser is very inefficient: it rechecks the entire buffer, even though when a new chunk arrives, we can guarantee that the regex cannot match the existing buffer, and it uses a capture that will always just capture everything from the beginning of the string to the point of a line break.

This PR addresses both of those deficiencies:

• When a new chunk arrives, only the new chunk is checked for a newline sequence.
• No captures are used; instead the index field of the result is used to pick the end of the slice sent for parsing.

[eclipse-cdt-bot]

Can one of the admins verify this patch?

[jonahgraham]

Its up npm: https://www.npmjs.com/package/cdt-gdb-adapter/v/0.0.16-next.20220908234947.bdf98e3.0

Thanks @colin-grant-work for the performance improvement.

[colin-grant-work]

Thanks, @jonahgraham!