Optimize outline parsing allocations and fast-path scalar properties

[tiensonqin]

Motivation

• Reduce allocations and work during outline parsing by reusing the strings produced by String.split_on_char instead of mapping and allocating new strings.
• Avoid computing line_starts when only an outline parse is requested via config.parse_outline_only.
• Avoid running the full inline grammar for property values that cannot contain references to improve performance on common scalar Logseq properties.

Description

• In lib/syntax/md_outline.ml replace Array.of_list (List.map rstrip_cr raw_lines) with Array.of_list raw_lines and Array.iteri to only replace CRLF lines in place, reusing the original allocations for LF inputs.
• Make line_starts allocation conditional on config.parse_outline_only so the array is not built when not needed.
• In lib/syntax/property.ml add a cheap may_have_reference scan for '#', '[', or '(' and return early when none are present, and keep the existing quoted-string check; only run the full Inline.parse when a reference is plausible.

Testing

• Ran the project test suite with dune runtest, and the tests completed successfully.
• Performed quick parsing smoke tests on sample outlines and property lines to verify no regressions in outline extraction and property reference detection.

---

Codex Task

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}