docs: record that main is reached by pr because push runs no ci #4

Merged

gravityfargo merged 1 commit from docs/record-pr-workflow into main

2026-08-02 22:49:32 +00:00

gravityfargo commented

2026-08-02 22:33:54 +00:00

Owner

.forgejo/workflows/test.yml triggers on workflow_call and pull_request only. There is no push trigger, so commits pushed straight to main are never linted, typechecked, or tested.

Every Actions run in this repo so far has event: pull_request, which confirms it — today's work on main carries no CI record.

The workflows are template-owned (only ever changed by chore: rebuild dev env from templates and ci: gate tests on code quality and sync the template baseline), so a push trigger belongs in $DEV_DIR/templates, not in a local edit here. This records the branch-and-PR convention instead.

Also notes uv lock --check, which code-quality.yml runs and make pre-commit does not — a stale lock passes uv sync --frozen silently.

This PR is itself the first use of the convention.

`.forgejo/workflows/test.yml` triggers on `workflow_call` and `pull_request` only. There is no `push` trigger, so commits pushed straight to `main` are never linted, typechecked, or tested. Every Actions run in this repo so far has `event: pull_request`, which confirms it — today's work on `main` carries no CI record. The workflows are template-owned (only ever changed by `chore: rebuild dev env from templates` and `ci: gate tests on code quality and sync the template baseline`), so a `push` trigger belongs in `$DEV_DIR/templates`, not in a local edit here. This records the branch-and-PR convention instead. Also notes `uv lock --check`, which `code-quality.yml` runs and `make pre-commit` does not — a stale lock passes `uv sync --frozen` silently. This PR is itself the first use of the convention.

gravityfargo added 1 commit

2026-08-02 22:33:54 +00:00

docs: record that main is reached by pr because push runs no ci b534090f18