Summary
The resolveCases function in monarchRunner.ts uses Array.includes() (O(n)) to check if a matched token belongs to a named category (keywords, BSON constructors, shell commands, operators). Converting these arrays to Set objects would make lookups O(1).
Context
From PR #580 review item I-05.
File: src/documentdb/shell/highlighting/monarchRunner.ts (line ~195)
if (Array.isArray(array) && (array as string[]).includes(matchedText)) {With 47 keywords, 10 BSON constructors, and 8 shell commands, the linear scan is called on every identifier token during tokenization. For typical shell input (< 200 chars), this is well within the 0.5ms target and is not a correctness issue.
Proposed Fix
Lazily convert the readonly string[] arrays in MonarchLanguageRules to Set<string> on first use (e.g., via a WeakMap<MonarchLanguageRules, Map<string, Set<string>>> cache), then use Set.has() instead of Array.includes().
Priority
Low — This is a minor performance optimization. Current performance is acceptable for all practical shell input sizes.
Summary
The
resolveCasesfunction inmonarchRunner.tsusesArray.includes()(O(n)) to check if a matched token belongs to a named category (keywords, BSON constructors, shell commands, operators). Converting these arrays toSetobjects would make lookups O(1).Context
From PR #580 review item I-05.
File:
src/documentdb/shell/highlighting/monarchRunner.ts(line ~195)With 47 keywords, 10 BSON constructors, and 8 shell commands, the linear scan is called on every identifier token during tokenization. For typical shell input (< 200 chars), this is well within the 0.5ms target and is not a correctness issue.
Proposed Fix
Lazily convert the
readonly string[]arrays inMonarchLanguageRulestoSet<string>on first use (e.g., via aWeakMap<MonarchLanguageRules, Map<string, Set<string>>>cache), then useSet.has()instead ofArray.includes().Priority
Low — This is a minor performance optimization. Current performance is acceptable for all practical shell input sizes.