feat(schema): surface all ResolutionOverrides fields in doctor --json#83
Merged
Conversation
OverridesV3 (now Overrides) only carried pm/runner/prefer_runners/ fallback/on_mismatch/explain/no_warnings/quiet — a field could land on ResolutionOverrides and CLI/env/config but never reach doctor --json, which is exactly the drift #77 flagged with quiet. Add the rest: - prefer_sources ([tasks].prefer) - task_source_pins ([tasks.overrides]; reported name avoids clippy's struct_field_names lint on task_source_overrides) - failure_policy (-k/-K / RUNNER_KEEP_GOING/RUNNER_KILL_ON_FAIL / [chain]) - output_grouping.{group_output,github_group_parallel,parallel_grouped} ([github].group_output/group_parallel, [parallel].grouped; nested to keep Overrides under clippy's struct_excessive_bools threshold) - install_pms (RUNNER_INSTALL_PMS / [install].pms) - script_policy (--no-scripts/--scripts / RUNNER_INSTALL_SCRIPTS / [install].scripts) parent_group_open is the one ResolutionOverrides field left out: internal runner-to-runner plumbing (an inherited env marker), never a user override. Added a drift guard test (every_resolution_overrides_field_is_reported_or_excluded) mirroring #77's KNOWN_SCHEMA-vs-RunnerConfig structural check: every ResolutionOverrides field must appear in Overrides or on its exclusion list, so a future field can't silently miss both again. Closes #81
This comment was marked as outdated.
This comment was marked as outdated.
kjanat
added a commit
that referenced
this pull request
Jul 5, 2026
render_init_template() panics by design on FIELD_TEMPLATE/RunnerConfig drift (a hard failure is right for the drift-guard test that normally catches this pre-merge) — but write_all_schemas called it unconditionally, so a released binary hitting that drift would crash `runner schema --all` with a raw, unhandled panic instead of a normal error exit. checked_init_template() wraps the call in catch_unwind with the default panic hook suppressed (no backtrace noise), converting a caught panic into an anyhow::Error with the same message. Sabotage-verified: an out-of-vocabulary FIELD_TEMPLATE value now exits 1 with a clean 'Error: ...' line instead of a panic backtrace. Not fixed here: the same review also flagged doctor.rs's Overrides struct missing prefer_sources/task_source_overrides/failure_policy/ group_output/github_group_parallel/parallel_grouped/install_pms/ script_policy/parent_group_open. That's real on this branch alone, but it's exactly issue #81's scope, already fully implemented (all 9 of those fields, plus a structural drift-guard test) on the stacked branch issue-81-overrides-fields (PR #83, based on this one) — duplicating it here would conflict when #83 rebases after this merges.
kjanat
changed the title
doctor --json
coderabbitai
Bot
added
area: cli
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
…fields # Conflicts: # src/schema/doctor.rs
Note parent_group_open's exclusion explicitly, add pm_by_ecosystem to the "previously surfaced" list (it predates this PR, isn't new). Skipped the schemas/doctor.schema.json required-list finding: quiet staying out of required is deliberate (remove_required_def_field, back-compat), already guarded by two tests.
remove_required_def_field/patch_schema_compat kept quiet optional for compat with doctor v3, which 2d397fe already deleted (schema-version now rejects anything but 1). No compat boundary left to preserve. Removed the shim and its two tests; quiet now required like every other Overrides field.
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
… source labels FailurePolicy::label() and ScriptPolicy::report_label() (Default included, unlike the settable-input label()) replace inline matches in overrides_report, matching fallback/on_mismatch's existing .label() use. Overrides.prefer_sources/task_source_pins now $ref TaskSourceLabel (same closed set SourceEntry.kind already uses) instead of plain strings. Skipped: turning failure_policy/script_policy into typed schema enums. Overrides has 6 other equally-closed-set label fields (fallback, on_mismatch, pm, runner, prefer_runners, install_pms) that stay plain strings by existing convention; enum-typing 2 of 8 would make the struct's own schema inconsistent rather than fix anything. A full switch is a separate, deliberate call, not a minimal patch.
…-built JSON Reworked the previous commit's approach. FailurePolicy, FallbackPolicy, MismatchPolicy, ScriptPolicy, PackageManager, TaskRunner now derive schemars::JsonSchema + Serialize (kebab-case, matching each .label()'s existing wire strings; TaskRunner::GoTask needs an explicit rename to "task" since kebab-case alone gives "go-task", the parse alias not the canonical label). Overrides' fields hold these enums directly instead of flattened &'static str/Option<&'static str>/Vec<&'static str>, so schemars derives each field's real oneOf/enum constraint automatically -- no hand-written json!() schema fragments, no patch_def_* plumbing for them. Removed the now-fully-redundant FailurePolicy::label/ALL and ScriptPolicy::ALL/ report_label added in the prior commit for that purpose. TaskSourceLabel-based patching stays for prefer_sources/task_source_pins -- those vary per command (doctor/why structured vs list flat), so they still need the runtime label-function-driven $ref, unlike the other fields which are simple closed enums schemars can derive standalone. Wire format (doctor --json's overrides object) is byte-identical; verified via doctor --json and the committed example fixture (no diff).
You were right to push on this -- schemars 1.2's BTreeMap<K, V> impl
does constrain keys automatically when K's schema is the flat
{"type":"string","enum":[...]} shape. It wasn't kicking in because
Ecosystem's per-variant /// doc comments made schemars emit oneOf+const
instead (schemars_derive treats any doc'd variant as "complex"). Doc
comments moved to // (kept, just not schema-visible) and Ecosystem
gained the same Serialize/JsonSchema/kebab-case treatment as the other
enums this session, plus Ord (needed as a BTreeMap key).
Overrides.pm_by_ecosystem now keys by Ecosystem directly; doctor.rs
constructs it without the .label().to_string() detour. Schema gets
explicit per-ecosystem properties + additionalProperties: false instead
of a bare object. Wire output unchanged (verified via doctor --json and
the committed example fixture).
coderabbitai
Bot
added
the
documentation
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
…ngelog wording every_resolution_overrides_field_is_reported_or_excluded now destructures ResolutionOverrides::default() with no `..` -- a new struct field fails compilation right there instead of the hand-maintained RESOLUTION_OVERRIDES_FIELDS list silently staying stale. Verified via a sabotage field (E0027 fires as expected), reverted. CHANGELOG: failure_policy/script_policy/install_pms are new fields from this PR, not pre-existing generic strings tightened into enums -- moved them out of that framing.
- drift guard: one macro list now both destructures ResolutionOverrides and feeds the checked names — the array can no longer go stale behind the compile-time check (sabotage-verified both directions) - pin pm_by_ecosystem's generated schema shape (7 closed keys, additionalProperties:false) so a stray /// on an Ecosystem variant can't silently reopen the key set - pin serde kebab-case output to label() for Ecosystem/PackageManager/ TaskRunner/FallbackPolicy/MismatchPolicy/ScriptPolicy (GoTask-style drift is now caught by test, not author vigilance) - Ecosystem: manual Ord by label — doctor pm_by_ecosystem keys stay alphabetical (pre-existing order, matches info/list surface) instead of flipping to declaration order - drop the never-None Option on pm_by_ecosystem values; schema loses the bogus null branch on all seven keys - schema.rs: field_schema_mut/def_ref consolidate the three patch fns' copy-pasted navigation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #81. Stacked on #82 (base branch
issue-80-schema-v1-purge) — therenamed
Overrides/DoctorReporttypes this PR touches only exist afterthat one merges.
OverridesV3(nowOverrides) only carriedpm/runner/prefer_runners/fallback/on_mismatch/explain/no_warnings/quiet— a field could landon
ResolutionOverridesand CLI/env/config but never reachdoctor --json,exactly the drift #77 flagged with
quiet. Adds the rest:prefer_sources([tasks].prefer)task_source_pins([tasks.overrides]; reported name avoids clippy'sstruct_field_nameslint ontask_source_overrides)failure_policy(-k/-K/RUNNER_KEEP_GOING/RUNNER_KILL_ON_FAIL/[chain])output_grouping.{group_output,github_group_parallel,parallel_grouped}(
[github].group_output/group_parallel,[parallel].grouped; nestedto keep
Overridesunder clippy'sstruct_excessive_boolsthreshold)install_pms(RUNNER_INSTALL_PMS/[install].pms)script_policy(--no-scripts/--scripts/RUNNER_INSTALL_SCRIPTS/[install].scripts)parent_group_openis the one field left out: internal runner-to-runnerplumbing (an inherited env marker), never a user override.
Adds a drift guard test
(
every_resolution_overrides_field_is_reported_or_excluded) mirroring#77's
KNOWN_SCHEMA-vs-RunnerConfigstructural check: everyResolutionOverridesfield must appear inOverridesor on its exclusionlist, so a future field can't silently miss both again.
Test plan
cargo test --all-features(881 unit + all integration suites green)cargo clippy --all-targets --all-features -- -D warnings -D clippy::allcleancargo fmt --checkcleancargo run --features schema -- schema --all --output schemasis idempotent against the committed files