From v3 onwards, the `gradle-build-action` will be soft deprecated in preference
of `gradle/actions/setup-gradle`, which is a drop-in replacement. Version numbers
for these 2 actions will remain in sync.
Workflows that use `gradle/gradle-build-action` will transparently delegate to
`gradle/actions/setup-gradle`. At a later date, a deprecation warning will be added
to `gradle/gradle-build-action`, encouraging users to migrate.
All of the repository sources have been migrated to the `gradle/actions` repository,
and `gradle/actions/setup-gradle` has been released with `v3.0.0-beta.6`.
Going forward, releases will be first performed for `gradle/actions/setup-gradle`,
with a parallel release of `gradle/gradle-build-action` soon afterward.
Releasing as `gradle/gradle-build-action@v3` will allow dependabot to encourage users
to upgrade from `v2`. The deprecation warning (to be added later) will encourage users
to migrate to `gradle/actions/setup-gradle`.
When changing workflow names or when changing to the new 'dependency-submission'
action, it can be useful to clear existing dependency graph snapshots from previous
submissions. While the old graphs will eventually "age out", the 'clear' option will
submit an empty dependency graph for an existing Job correlator, ensuring that old
dependency graphs don't linger.
- Translate to env var for init-script support
- Use when deciding whether to log or rethrow errors
- Add a custom error type to trigger failure in post action
Instead of a binary true/false option, it is now possible to only add
a Job Summary when the build failed. This applies both to the overall
Job Summary added to the workflow run, and to the new PR comment feature.
Rather than requiring a separate step to add a PR comment,
the `gradle-build-action` can now automatically add the Job Summary
as a PR comment
Fixes#1020
- Don't upload artifacts when using 'generate-and-submit'
- New option 'generate-and-upload' to be used with 'download-and-submit'
- Use Artifact API for downloading in the same and different workflow
This makes it easier for users to enable config-cache saving in their workflow.
Config-cache data will only be saved/restored when the key is provided,
and the key is exported as `GRADLE_ENCRYPTION_KEY` for use in subsequent steps.
- Added a new `artifact-retention-days` input parameter to control retention of uploaded artifacts
- Artifacts retention will use repository settings if not overridden.
Adds a 'dependency-graph' parameter that has 4 options:
1. 'disabled': no dependency graph files generated (the default)
2. 'generate': dependency graph files will be generated and saved as artifacts.
3. 'generate-and-submit': dependency graph files will be generated, saved as artifacts,
and submitted to the Dependency Submission API on job completion.
4. 'download-and-submit': any previously uploaded dependency graph artifacts will be downloaded
and submitted to the Dependency Submission API.
Instead of requiring an action step to generate the graph, configure Gradle User Home
so that subsequent Gradle invocations can generate a graph. Any generated graph files
are uploaded as artifacts on job completion.
- Construct job.correlator from workflow/job/matrix
- Export job.correlator as an environment var
- Upload artifacts at job completion in post-action step
- Specify the location of dependency graph report
- Only apply dependency graph init script when explicitly enabled
Introducing new actions for the GitHub dependency graph will involve reuse of much of
the action infrastructure. This commit reorganises things a little to facilitate reuse.
- Includes basic implementation as `CacheCleaner`
- Integration test that checks unused files are removed:
- Downloaded dependencies
- Local build cache entries
- Wrapper distributions
Now that `@actions/cache` provides a env var override, we can remove our
custom configuration for setting the timeout on cache restore operations.
If the env var is NOT set, we continue to override the 60min default with 10mins.
The fix for #383 introduced a bug that caused the post-action to fail when
attemping to resolve `cache-read-only` for scheduled jobs.
This has been fixed with an explicit (rather than implicit) check for 'event.repository != null'.
Fixes#409
For scheduled workflows, we don't know the default branch. In this case cannot determine if
the workflow is running on the default branch, and so cannot know when to enable `cache-read-only`.
Fixes#383
Cache entries _written_ from jobs run on a non-default branch will be private
to other jobs for that branch. When development flow involves working on a
feature branch and then merging into 'main', these branch-private cache
entries can result in eviction of other (shared) cache entries generated
for the default branch.
With this change, we make the recommended setup the default, by running
with `cache-read-only: true` for any jobs run on a non-default branch.
These jobs will be able to read cache entries written from the main branch,
but will not write any cache entries.
Fixes#143
There may be cases where it a "fresh" cache entry would be beneficial,
for example if the Gradle User Home cache entry grows over time.
This setting would run the build as if no prior cache entry exists.
By default, the action will attempt to restore a Gradle User Home
cache entry from a different set of matrix inputs (or a different Job entirely)
if an existing entry is not found for the current Job (including matrix inputs).
By specifying the experimental `gradle-home-cache-strict-match` parameter, a user
can avoid this fuzzy matching and ensure that a job execution starts with an
empty Gradle User Home if no entry from a prior execution is found.
Instead of using a fallback strategy to locate a configuration-cache entry
based on the current job and git SHA, these entries are now keyed based on their
file content with the keys persisted in the primary Gradle User Home entry.
This removes the chance of having a configuration-cache entry restored that is
incompatible with the restored Gradle User Home state, and makes the logic easier
to understand.
This change involved a fairly major refactor, with the CacheEntryExtractor being
split out from the primary cache implementation, and adding a separate extractor
implementation for configuration-cache.
- Remove configuration parameter for cache-entry-definitions. This allows richer modelling of extracted entries.
- Include both *.zip and *.tar.gz in entry definition for java-toolchains.
