* Return early from dry run
Determining whether to create a merge commit would elicit a nested
conditional, which could be hard to parse for a human reader. This is
avoided by returning early as soon as possible for a dry run.
This also resolves the erroneous 'changes committed' message when no
changes were actually committed because of the dry run. A message
specific to dry-run is logged instead.
* Add force parameter to action interface
Existing behaviour is equivalent to force=true, so the default value is
true.
* Implement pull+rebase procedure
* Declare force parameter in action
* Detect both rejection syntaxes
* Return both stdout and stderr from execute
* Ignore non-zero exit status on push
* Remove unnecessary error catch
* Fetch and rebase in separate steps
* Explicitly bind incoming branch
I think the fetch will update the origin/gh-pages branch but not the
gh-pages branch, despite requesting gh-pages. This means that when I
later attempt to rebase the temp branch on top of the gh-pages branch,
there will be nothing to do, because that's already where it is.
* Implement attempt limit
I don't expect this to ever require more than one attempt in production,
but in theory it's possible that this procedure could loop forever.
We would need to keep fetching and rebasing if changes keep being added
to the remote. In practice, I believe this would only happen if there
are lots of workflows simultaneously deploying to the same branch, all
using this action. In this case only one would be able to secure a lock
at a time, leading to the total number of attempts being equal to the
number of simultaneous deployments, assuming each deployment makes each
attempt at the exact same time.
The limit may need to be increased or even be configurable, but 3 should
cover most uses.
* Update tests for execute output split
* Document 'force' parameter
* Create integration test for rebase procedure
This test is composed of 3 jobs.
The first two jobs run simultaneously, and as such both depend on the
previous integration test only. The final job cleans up afterwards, and
depends on both of the prior jobs.
The two jobs are identical except that they both create a temporary file
in a different location. This is to ensure that they conflict. Correctly
resolving this conflict by rebasing one deployment over the other,
resulting in a deployment containing both files, indicates a successful
test.
GitHub reports when workflow runs log messages with `notice` or higher
logging levels, e.g.:
> There are 0 failures, 0 warnings, and 1 notices.
Since `notice` was being used regardless of status, these reports were
misleading, because everything was working correctly on successes and
no-ops.
Therefore, the successes and no-ops now only log with `info`, and the
failures now only log with `notice`.
* Stop checking out workspace (#515)
* Stop checking out base branch before deployment, drop option.
* Don't check out default branch, as we don't check out base branch, drop option.
* Don't stash/unstash as we don't update the workdir, drop preserve option.
* Don't init the workspace
* Only fetch the remote branch if it exists, only with depth 1.
* Rely on previous checkouts to have handled lfs files correctly, drop option.
* Update README, action.yml, integration tests
* Set up eslint for test files. (#517)
* Add DRY_RUN option, passing --dry-run to git push. (#526)
See #499 for the proposal.
* Simplifies Token Setup (#530)
* Token simplification
* Access Token / Github Token -> Token
* Oops
* Typos
* Update README.md
* Update README.md
* Update action.yml
Co-authored-by: Axel Hecht <axel@pike.org>
* Update README.md
Co-authored-by: Axel Hecht <axel@pike.org>
* Update README.md
Co-authored-by: Axel Hecht <axel@pike.org>
* Adjust codeql action to latest recommendations (#540)
Also, add the dev and release branches, and drop master.
* Add workflow to update build and node_modules on release branches (#541)
* Stores username/email in secrets
* Removing stale bot integration
* Test current code base as an integration test for PRs and pushes (#505)
* Add a build step to create lib and node_modules artifact
* Run integration test with built dist and current SHA as base
For pull requests, the github.sha is the sha of the merge to the
target branch, not the head of the PR. Special case that.
* Use v2 checkout, and DRY_RUN for the integration test.
I also made the branches more generic, as there are now more of them.
* Fix#536, don't push at all on dryRun
Also add tests for dryRun and singleCommit and generateBranch
code flows.
* Try to fix dryRun on new remote branches, refactor fetch
* Try to fix dryRun, only fetch if origin branch exists
* Refactor worktree setup to include branch generation and setup for singleCommit
This is a continuation of the no-checkout work, and sadly suggested pretty
intensive changes.
* Set up git config to fix tests, also make debugging easier
* Add matrix for existing and non-existing branch
* Add matrix for singleCommit and not
* Drop GITHUB_TOKEN, add DRY_RUN to action.yml
* When deploying existing branch, add a modifcation and deploy again
* Force branch checkout to work in redeployment scenarios
* Make singleCommit easier to see in job descriptions
* Review comments
* Add a test-only property to action to test code paths with remote branch.
* Introduce TestFlag enum to signal different test scenarios to unit tests
* Fix util.test.ts
* Update worktree.ts
* Fix a few nits in tests and automation. Don't try to wordcount ls-rem… (#546)
* Fix a few nits in tests and automation. Don't try to wordcount ls-remote.
Nits in tests are around undoing changes made to the environment,
and to not modify the checkout.
* Describe suite with empty SHA
* Lowercase Inputs (#547)
* Lowercases inputs
* Adjusts workflow tests and deployment_status
* Use multi-line string for clean-exclude patterns. (#553)
As this change is subtle, I'm taking the opportunity to change
the underscore for the hyphen, which makes it less likely that
users of this action will just pass in an old json array.
* Hyphenate inputs and outputs, add step output, fix#558 (#559)
* Hyphenate inputs and outputs, add step output, fix#558
I've also tried to make the clean docs a bit clearer, and consistent
about clean being on my default. Still not totally happy with the intro
of the docs there, though.
* Add testing of step outputs to build integration tests
* Security Docs
* Integration tests
* Revert "Integration tests"
This reverts commit 639ff537d5.
* Native SSH Key Support (#569)
* SSH Key Support 🔑
* Update ssh.ts
* Update src/ssh.ts
Co-authored-by: Axel Hecht <axel@pike.org>
* README fixes/etc
* Unit Tests & README
* ssh key
* Update README.md
* Update ssh.test.ts
* Update ssh.test.ts
* Update ssh.test.ts
* Update ssh.test.ts
* Update ssh.test.ts
* Update ssh.test.ts
* Update integration.yml
Co-authored-by: Axel Hecht <axel@pike.org>
* Deployment Issues (#583)
* Update git.ts
* Tests
* Update git.ts
* Formatting
* Update src/git.ts
Co-authored-by: Axel Hecht <axel@pike.org>
* TestFlag
* Logging
* Update git.ts
Co-authored-by: Axel Hecht <axel@pike.org>
* Codespace Support (#584)
* Add files via upload
* Update README.md
* Add files via upload
* Update README.md
* Update README.md
* Update README.md
* Update README.md
* Update README.md
* SSH Issues (#588)
* Unsets Persisted Credentials (#587)
* Persist
* Config Setup/Tests
* Assets
* Update git.ts
* Spacing
* Update integration.yml
* Update README.md
Co-authored-by: Axel Hecht <axel@pike.org>
* Feat: add support for absolute paths
* Refactor: `generateFolderPath` to generic function
`action.root` is initialized at e42fda2d72/src/constants.ts (L98), so it should not be an optional property.
* Test: rm unnecessary tests
* Refactor: rm unused exports
