bench context

bench-context

Benchmarks the context command by building context bundles for a sample of symbols against a database. Reach for it to measure the cost of generating context bundles, not to inspect any single symbol.

Running it from the CLI

drag-lint bench-context [--db <file.sqlite>] [--n N]

--n N sets how many symbols the benchmark samples; --db selects the index. Neither flag's default is documented in the usage line.

Reaching it in the IDE

No IDE surface -- this is a CLI-only feature.

What it needs

The --db flag is optional -- drag-lint auto-resolves the index when omitted. An index must still exist.

Example

Illustrative only:

drag-lint bench-context --db C:\Projects\MyApp\_D-RAG\MyApp.sqlite --n 20

This would build context bundles for 20 sampled symbols and report benchmark results.