.pre-commit-config.yaml sets manual: true on two hooks: oca-gen-addon-readme (line 70) and oca-gen-external-dependencies (line 72).
manual is not a pre-commit hook key — pre-commit expects stages: [manual]. Given an unknown key it warns and ignores it, so both hooks keep running in the default stage.
The visible effect: any PR touching a readme/*.md fragment fails the pre-commit job until the regenerated README.rst and static/description/index.html are committed. That may well be the behaviour you want — but it isn't what the config says, and contributors hit it with no obvious explanation.
Two ways to resolve:
If the generators are meant to be manual, use stages: [manual] on both. CI then stops regenerating READMEs and someone runs them before release.
If they are meant to run every time (what happens today), drop the manual: true lines so the config states the actual behaviour, and note in the contributing docs that fragment changes require committing the regenerated files.
Noticed while addressing review on OpenSPP/OpenSPP2#397.
.pre-commit-config.yaml sets manual: true on two hooks: oca-gen-addon-readme (line 70) and oca-gen-external-dependencies (line 72).
manual is not a pre-commit hook key — pre-commit expects stages: [manual]. Given an unknown key it warns and ignores it, so both hooks keep running in the default stage.
The visible effect: any PR touching a readme/*.md fragment fails the pre-commit job until the regenerated README.rst and static/description/index.html are committed. That may well be the behaviour you want — but it isn't what the config says, and contributors hit it with no obvious explanation.
Two ways to resolve:
If the generators are meant to be manual, use stages: [manual] on both. CI then stops regenerating READMEs and someone runs them before release.
If they are meant to run every time (what happens today), drop the manual: true lines so the config states the actual behaviour, and note in the contributing docs that fragment changes require committing the regenerated files.
Noticed while addressing review on OpenSPP/OpenSPP2#397.