You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
rocky doctor gained a --verbose flag in #21 to surface extra detail (config paths, sizes, adapter types) in the text output path. The same pattern should apply to rocky estimate: a user diagnosing a cost estimate often wants the per-model EXPLAIN-plan raw output, which isn't currently surfaced.
Scope
Add --verbose to the Estimate command in engine/rocky/src/main.rs, thread it through run_estimate, and have the text output path print additional per-model detail (raw optimizer / EXPLAIN output, or extra cost breakdown — whatever the function already has on hand but doesn't currently print).
Out of scope: the JSON output. Do not change any *Output struct in rocky-cli/src/output.rs — that path cascades to Pydantic/TypeScript via just codegen. This issue is text-output-only.
Code refs
Line numbers refreshed against main 2026-06-14.
engine/rocky/src/main.rs:1322 — Estimate { ... } enum variant (add the verbose: bool field here, mirroring Doctor)
engine/rocky/src/main.rs:2989 — Command::Estimate { ... } dispatch arm and the run_estimate(...) call site
engine/crates/rocky-cli/src/commands/estimate.rs:22 — run_estimate function signature and text output path
Closed Add --verbose flag to rocky doctor #21 for the rocky doctor --verbose pattern to mirror (the Doctor { check, verbose } variant + dispatch is the template)
Acceptance criteria
rocky estimate --verbose prints additional per-model detail in the text output
rocky estimate (no flag) behavior is unchanged
The JSON output is unchanged (no *Output struct edits)
--help text for the flag is consistent with rocky doctor --verbose
rocky doctorgained a--verboseflag in #21 to surface extra detail (config paths, sizes, adapter types) in the text output path. The same pattern should apply torocky estimate: a user diagnosing a cost estimate often wants the per-model EXPLAIN-plan raw output, which isn't currently surfaced.Scope
Add
--verboseto theEstimatecommand inengine/rocky/src/main.rs, thread it throughrun_estimate, and have the text output path print additional per-model detail (raw optimizer / EXPLAIN output, or extra cost breakdown — whatever the function already has on hand but doesn't currently print).Out of scope: the JSON output. Do not change any
*Outputstruct inrocky-cli/src/output.rs— that path cascades to Pydantic/TypeScript viajust codegen. This issue is text-output-only.Code refs
engine/rocky/src/main.rs:1322—Estimate { ... }enum variant (add theverbose: boolfield here, mirroringDoctor)engine/rocky/src/main.rs:2989—Command::Estimate { ... }dispatch arm and therun_estimate(...)call siteengine/crates/rocky-cli/src/commands/estimate.rs:22—run_estimatefunction signature and text output pathrocky doctor --verbosepattern to mirror (theDoctor { check, verbose }variant + dispatch is the template)Acceptance criteria
rocky estimate --verboseprints additional per-model detail in the text outputrocky estimate(no flag) behavior is unchanged*Outputstruct edits)--helptext for the flag is consistent withrocky doctor --verbosecargo test -p rocky-cliis greenEffort
2 hours.