perf(parser): cache has_separator() result in literal parsing

[Boshen]

Summary

Cache the token.has_separator() result in parse_literal_number() and parse_literal_bigint() to avoid redundant memory reads on the hot path.

[Copilot]

Pull Request Overview

Optimize numeric literal parsing by caching the has_separator() result to eliminate redundant memory reads and pointer arithmetic operations.

• Cache token.has_separator() result in local variable instead of calling it multiple times
• Apply optimization to both parse_literal_number() and parse_literal_bigint() functions
• Reduce memory access overhead on the hot path for parsing millions of numeric literals

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[graphite-app]

How to use the Graphite Merge Queue

Add either label to this PR to merge it via the merge queue:

• 0-merge - adds this PR to the back of the merge queue
• hotfix - for urgent hot fixes, skip the queue and merge this PR next

You must have a Graphite account in order to use the merge queue. Sign up using this link.

An organization admin has enabled the Graphite Merge Queue in this repository.

Please do not merge from GitHub as this will restart CI on PRs being processed by the merge queue.

[Boshen]

Merge activity

• Oct 7, 9:10 AM UTC: The merge label '0-merge' was detected. This PR will be added to the Graphite merge queue once it meets the requirements.
• Oct 7, 9:13 AM UTC: Boshen added this pull request to the Graphite merge queue.
• Oct 7, 9:18 AM UTC: The Graphite merge queue couldn't merge this PR because it was not satisfying all requirements (Failed CI: 'Test NAPI').

[codspeed-hq]

CodSpeed Performance Report

Merging #14403 will not alter performance

_{Comparing perf/cache-has-separator-in-literal-parsing (29fa3a9) with main (bce31b5)}

Summary

✅ 37 untouched