@aurodesignsystem-dev/auro-cli 0.0.0-pr350.0 → 0.0.0-pr350.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.md CHANGED
@@ -215,7 +215,7 @@ Only public API counts. Members are filtered on the manifest's `privacy` field,
215
215
 
216
216
  The manifest is rebuilt at both refs in temporary git worktrees, by default with `npm ci` and `npm run build`. The rebuild is skipped when only `test/`, `.github/` or `docs/` changed. If the head doesn't build, the check errors (fails closed). If the base doesn't build, the check uses the base's committed manifest, or skips the API comparison, and says so in `warnings` (fails open): a PR shouldn't be blocked by a base it didn't break.
217
217
 
218
- **Declared bump.** The highest bump among the non-merge commits, worked out by `@semantic-release/commit-analyzer` with the repo's own release config (usually `@aurodesignsystem/auro-config/releaserc`, including its custom rules: `refactor`, `style` and `docs(README)` release a patch). It's loaded from the repo's `node_modules` when installed, so the check reads commits exactly as the release will. Otherwise auro-cli's own copy and the shared rules are used. The PR title never counts, because merge commits don't affect the version.
218
+ **Declared bump.** The highest bump among the non-merge commits, worked out by `@semantic-release/commit-analyzer` with the repo's own release config (usually `@aurodesignsystem/auro-config/releaserc`, including its custom rules: `refactor`, `style` and `docs(README)` release a patch). The analyzer is loaded from the repo's own `node_modules`, so the check reads commits exactly as the release will. Run `npm ci` in the repo first: the check errors when the analyzer isn't installed. If the repo's release config can't be resolved, the shared rules are used and `warnings` says so. The PR title never counts, because merge commits don't affect the version.
219
219
 
220
220
  **Outcome.**
221
221
 
@@ -223,7 +223,7 @@ The manifest is rebuilt at both refs in temporary git worktrees, by default with
223
223
  - `warn`: no release is required, but the commits declare one (for example, a `fix:` that only touches tests).
224
224
  - `pass`: otherwise. Declaring more than required is allowed, because behavior can break without the API changing.
225
225
 
226
- The JSON starts with `required` and `reasons`, followed by `declared`, `outcome`, a one-line `summary` for PR comments, the refs, each change and commit with its bump, the release config and analyzer used, and any `warnings`:
226
+ The JSON starts with `required` and `reasons`, followed by `declared`, `outcome`, a one-line `summary` for PR comments, the refs, each change and commit with its bump, the release config used, and any `warnings`:
227
227
 
228
228
  ```json
229
229
  {
@@ -235,7 +235,7 @@ The JSON starts with `required` and `reasons`, followed by `declared`, `outcome`
235
235
  }
236
236
  ```
237
237
 
238
- Exit codes: `0` when the analysis ran, `1` with `--enforce` when the outcome is `fail`, `2` when the analysis couldn't run (an unknown ref, or a head that won't build).
238
+ Exit codes: `0` when the analysis ran, `1` with `--enforce` when the outcome is `fail`, `2` when the analysis couldn't run (an unknown ref, a head that won't build, or no commit-analyzer installed in the repo).
239
239
 
240
240
  ### AI assistant context
241
241