perf(lsp): skip import lexing outside strings

[0xKarl98]

Towards OSS-738 ,

import_path_at handled every definition lookup by running the lexer from the beginning of the source file. Most requests come from ordinary code, where an import string is impossible.

• Reuse the existing may_complete_string scan before invoking plain_string_at.
• Return immediately when the cursor line has no quote and no escaped line continuation.
• Keep the existing lexer and parser path for strings, comments containing quotes, and multiline continuations.

Benchmark evidence (cargo bench -p solar-lsp --features bench --bench lsp -- --noplot):

Case	Before	After	Change
lsp/import-path/optimism-code	5.40–5.45 ms	3.36–3.42 ns	~1.6M× faster

The benchmark uses a 5.38 MB source with the cursor in ordinary code near the end of the file. Criterion compared a 20-sample baseline with a 30-sample optimized run.

[codspeed-hq]

Merging this PR will improve performance by ×31,000

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 1 improved benchmark
✅ 230 untouched benchmarks
⏩ 18 skipped benchmarks¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	lsp/import-path[optimism-code]	59,084.2 µs	1.9 µs	×31,000

Tip

Curious why performance improved? Comment @codspeedbot explain why performance improved on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing 0xKarl98:perf/lsp-completion-prefix (441e4b4) with main (b4727b7)}

Footnotes

1. 18 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[mattsse]

lgtm