feat: 14 - dry run

[joefrost01]

What problem are you trying to solve?

--dry-run only printed a one-line count summary and did not provide the execution plan details required by the spec (resolved files, format summary, SQL/config overview). It also errored for zero matched files, which made dry-run less useful as an informational planning mode.

What does this PR change?

This implements a detailed dry-run plan renderer in the query pipeline, including file list preview, format/schema/query/output summary, and key execution flags. It also adds dry-run-specific file resolution behavior for zero-file scenarios and --pipe data placeholder handling without consuming stdin payload.

Does this change align with DESIGN.md?

Yes. Dry run remains an early pipeline exit and still avoids DuckDB execution and downstream processing. The change only expands pre-execution reporting and keeps normal query behavior unchanged when --dry-run is not set.

What alternatives did you consider?

I considered extending the existing single-line dry-run message with ad-hoc fields, but that made testing and formatting brittle. Building a dedicated plan string generator made behavior explicit and testable, including truncation/mixed-format logic.

Does this PR contain multiple unrelated changes?

No. All changes are scoped to feature 14 dry-run behavior.

Existing PRs

• I have reviewed all open AND closed PRs for duplicates or prior art
• Related PRs: none found

Testing

• cargo test passes
• cargo clippy passes with no warnings
• cargo fmt has been run
• New tests added:
  • dry_run_plan_truncates_file_list_and_shows_summary_fields
  • dry_run_allows_zero_files_for_missing_glob
  • dry_run_pipe_data_uses_stdin_placeholder_without_reading_input

Evaluation

• What was the specific scenario you tested?
  • Dry run with many resolved files (list truncation + summary fields)
  • Dry run with no glob matches (informational success)
  • Dry run with --pipe data and stdin format set (stdin placeholder plan path)
• What was the output before and after the change?
  • Before: one-line dry-run summary only; zero-file dry-runs errored.
  • After: full multi-field dry-run plan to stderr with bounded file preview and zero-file informational output.
• Did you test error cases (bad input, missing files, invalid SQL)?
  • Yes, existing suite continues to cover non-dry-run error paths; dry-run-specific zero-file and pipe-data cases are now covered by new tests.

Human review

• A human has reviewed the COMPLETE proposed diff before submission

[joefrost01]

Claude review complete: LGTM on dry-run behavior, test coverage, and code quality; ready to merge once CI passes.