Tests involving Docker are skipped when `runtime.GOOS` is not `linux`. That is unnecessarily rigid and hinders development. There are other operating systems that can run Docker. It makes it impossible to run integration tests on Linux without Docker. Verifying that all tests that require Docker are skipped without Docker being present is impossible on Linux due to that, too.
Instead of relying on `runtime.GOOS`, developers can control whether tests involving Docker should be run with the help of the environment variable `TEST_FEATURES`.
If the variable is undefined, tests involving Docker will be run, because they are enabled by default:
```
$ make integration-test
```
That is the same as:
```
$ TEST_FEATURES="docker" make integration-test
```
If the variable does not contain the string `docker`, tests involving Docker will be **skipped**:
```
$ TEST_FEATURES="" make integration-test
```
If a test requires Docker, use `testutils.RequireTestFeatures(t, testutils.TestFeatureDocker)` to indicate that.
<!--start release-notes-assistant-->
<!--URL:https://code.forgejo.org/forgejo/runner-->
- other
- [PR](https://code.forgejo.org/forgejo/runner/pulls/1585): <!--number 1585 --><!--line 0 --><!--description dGVzdDogdXNlIGVudmlyb25tZW50IHZhcmlhYmxlIGZvciBza2lwcGluZyBEb2NrZXIgdGVzdHM=-->test: use environment variable for skipping Docker tests<!--description-->
<!--end release-notes-assistant-->
Reviewed-on: https://code.forgejo.org/forgejo/runner/pulls/1585
Reviewed-by: Mathieu Fenniak <mfenniak@noreply.code.forgejo.org>
#1563 made changes to the `exprparser` library to allow detecting access to `env.*` and `secrets.*`. Two bugs were discovered during failed refactoring efforts (#1566):
- `env.*` and `secrets.*` weren't detected as access; in the event that these expressions were used in an `if` block they wouldn't have caused Forgejo to defer execution to the runner rather than doing server-side `if` evaluation
- `toJSON(matrix)` outputs a wrapping `"Matrix"` element, which was missed when dynamic `runs-on` capabilities were added, but missed until tests triggered similar problems in `env`
More comprehensive unit test cases around both are fixed here. Practically speaking neither bug is likely to occur in reality.
<!--start release-notes-assistant-->
<!--URL:https://code.forgejo.org/forgejo/runner-->
- bug fixes
- [PR](https://code.forgejo.org/forgejo/runner/pulls/1572): <!--number 1572 --><!--line 0 --><!--description Zml4OiBleHByZXNzaW9uIGV2YWx1YXRpb24gZml4ZXMgZGlzY292ZXJlZCB3aGlsZSByZWZhY3RvcmluZyAnaWYnIGRldGVjdGlvbg==-->fix: expression evaluation fixes discovered while refactoring 'if' detection<!--description-->
<!--end release-notes-assistant-->
Reviewed-on: https://code.forgejo.org/forgejo/runner/pulls/1572
Reviewed-by: Andreas Ahlenstorf <aahlenst@noreply.code.forgejo.org>
This PR contains two fixes to the production code:
- 20d56ace53adf26f66e7eb14f3f899087aec70eb fixes the `hashFile` function in an action misinterpreting paths on Windows
- 4e2609329dc2e3e630a67c3237e38a92dd7dc784 fixes the `validate` command outputting full paths due to OS-unspecific path trimming logic
The remaining changes adopt tests to properly deal with
- Docker with Linux containers not being available on standard Windows runners c492fe121e2f2d2eb6dbc8fd5b1cef543437fa99
- Windows paths (6781d823683428d5cf8b9bac4753a458882b5994), (8ca14d0f4d4274e808fb027e021bac701279381e), (eb70aca72c89c08cdc03c35dd26f3e57bdeaacac), (4e2609329dc2e3e630a67c3237e38a92dd7dc784), (c00d4e0a81634c098086b822884eb6b525ebe933)
- Revert a change causing a test-runner panic on Windows (6608fec705f7e7ed48c7ffed3839f2b4c4c1defb)
- Exclude the daemon suicide test because it cannot be properly observed on Windows (3ef33b93b6c3422ce47325bc5330d2d954998cf4)
or in one case separate dedicated Windows and Linux test setups (709a1d3fdbfc8a6149a57fcf00570e842ef4d040)
I tried to keep the changes as minimal as possible and focussed on getting the tests to run again.
I only added skips for tests that I could not get to run due to the Docker requirements, or the different shutdown behaviors on Linux and Windows.
In the future, I will investigate why the defensive `chdir` in `linux_container_environment_extensions_test.go` (6608fec705f7e7ed48c7ffed3839f2b4c4c1defb) is necessary and whether I can get all of the docker tests (c492fe121e2f2d2eb6dbc8fd5b1cef543437fa99) to run on GitHub Actions
The passing of all tests on windows has been confirmed [here](https://github.com/Crown0815/Forgejo-runner-windows-builder/actions/runs/19344650459).
<!--start release-notes-assistant-->
<!--URL:https://code.forgejo.org/forgejo/runner-->
- other
- [PR](https://code.forgejo.org/forgejo/runner/pulls/1152): <!--number 1152 --><!--line 0 --><!--description V2luZG93cy1zcGVjaWZpYy1maXhlcw==-->Windows-specific-fixes<!--description-->
<!--end release-notes-assistant-->
Co-authored-by: Felix Kröner <felix.kroener@bruker.com>
Reviewed-on: https://code.forgejo.org/forgejo/runner/pulls/1152
Reviewed-by: Michael Kriese <michael.kriese@gmx.de>
Reviewed-by: Mathieu Fenniak <mfenniak@noreply.code.forgejo.org>
Co-authored-by: Crown0815 <crown0815@noreply.code.forgejo.org>
Co-committed-by: Crown0815 <crown0815@noreply.code.forgejo.org>
* feat: add post step to actions and add state command
This commit includes requried changes for running post steps
for local and remote actions.
This allows general cleanup work to be done after executing
an action.
Communication is allowed between this steps, by using the
action state.
* feat: collect pre and post steps for composite actions
* refactor: move composite action logic into own file
* refactor: restructure composite handling
* feat: run composite post steps during post step lifecycle
* refactor: remove duplicate log output
* feat: run all composite post actions in a step
Since composite actions could have multiple pre/post steps inside,
we need to run all of them in a single top-level pre/post step.
This PR includes a test case for this and the correct order of steps
to be executed.
* refactor: remove unused lines of code
* refactor: simplify test expression
* fix: use composite job logger
* fix: make step output more readable
* fix: enforce running all post executor
To make sure every post executor/step is executed, it is chained
with it's own Finally executor.
* fix: do not run post step if no step result is available
Having no step result means we do not run any step (neither pre
nor main) and we do not need to run post.
* fix: setup defaults
If no pre-if or post-if is given, it should default to 'always()'.
This could be set even if there is no pre or post step.
In fact this is required for composite actions and included post
steps to run.
* fix: output step related if expression
* test: update expectation
* feat: run pre step from actions (#1110)
This PR implements running pre steps for remote actions.
This includes remote actions using inside local composite actions.
* fix: set correct expr default status checks
For post-if conditions the default status check should be
always(), while for all other if expression the default status
check is success()
References:
https://docs.github.com/en/actions/learn-github-actions/expressions#status-check-functionshttps://docs.github.com/en/actions/creating-actions/metadata-syntax-for-github-actions#runspost-if
* fix: remove code added during rebase
