* artifactPrepareVersion: additional options
it is now possible to use
* a Unix timestamp number
* a short commit hash
* Add dedicated Docker versioning scheme
Avoid maven error `Unknown lifecycle phase \"-\"` when the value of a define contains `-`.
Don't split and trim maven arguments. Expect they come in as a list, keep them as list.
This is a breaking change compared to the old Groovy implementation which relied on using a shell for calling maven.
As an example, consider this diff:
```diff
- goals: 'org.apache.maven.plugins:maven-help-plugin:3.1.0:evaluate',
- defines: "-Dexpression=$pomPathExpression -DforceStdout -q",
+ goals: ['org.apache.maven.plugins:maven-help-plugin:3.1.0:evaluate'],
+ defines: ["-Dexpression=$pomPathExpression", "-DforceStdout", "-q"],
```
* fix docu
* enhanced step docu and metadata
* regenerate after changing metadata
* small wording change
Co-authored-by: Marcus Holl <marcus.holl@sap.com>
Do not exit with os.Exit(1) but using log.Entry().Fatal() instead
* Golang: forward error details
* extend groovy wrapper to provide proper error message
* create closure for error handling
This change enables the use of the artifactPrepareVersion go step instead
of the deprecated artifactSetVersion groovy implementation for the Cloud
SDK Pipeline. The groovy wrapper for the artifactPrepareVersion step now
enables the use of the download cache and minor issues are fixed.
With the step gctsCreateRepository it is possible to create a local gCTS repository on an ABAP server
Co-authored-by: Marcus Holl <marcus.holl@sap.com>
Also establish mavenExecute() via new JenkinsMavenExecuteRule in Groovy Unit Tests.
Co-authored-by: Stephan Aßmus <stephan.assmus@sap.com>
Co-authored-by: Christopher Fenner <26137398+CCFenner@users.noreply.github.com>
* Extends kubernetesDeploy step to support Helm 3
Currently, the kubernetesDeploy step has no support to Helm 3 due to the fact that:
- the initialization command used works only for Helm 2
- the image used when running the helm CLI is based on Helm 2
The need for Helm 3 support comes from the fact that Helm 3 introduces major architectural changes,
more specifically, the removal of its server-side agent called Tiller - thus, being incompatible with
one another.
This commit adds this support by introducing a new configuration field (helmVersion).
By default, its values is set to 2 (Helm 2) to avoid breaking any existing functionalities.
* Use deployTool field to decide between Helm 2 or 3
* Remove helm init and replace wait for atomic in v3
* Update cmd/kubernetesDeploy.go
Nice catch!
Co-Authored-By: Christopher Fenner <26137398+CCFenner@users.noreply.github.com>
* Add documentation for kubernetesDeploy step
* Add helm3 example for kubernetesDeploy step using mandatory fields
* Add new line at the end of kubernetesDeploy documentation
* Link kubernetesDeploy step with docs generator
* Add possible values for deployTool in kubernetesDeploy
* dummy change
* Revert "dummy change"
Co-authored-by: Oliver Nocon <33484802+OliverNocon@users.noreply.github.com>
Co-authored-by: Christopher Fenner <26137398+CCFenner@users.noreply.github.com>
* sonar: fix test cases
some unit tests uses path.Joint() instead of filepath.Join()
as the code does. on window these unit tests fail.
* missing newline
* Update cmd/sonarExecuteScan_test.go
Co-authored-by: Christopher Fenner <26137398+CCFenner@users.noreply.github.com>
Changed from passing config as an array that was implemented badly to correct implementation that JSON config gets passed with ' '
Co-authored-by: Oliver Nocon <33484802+OliverNocon@users.noreply.github.com>
* use sonar go in groovy
* use SONAR_TOKEN
* only use owner/repo if both are set
* trim version to major version digit
* fix code climate issues
* remove Sonar defaults
* use sonar go step
* use SONAR_TOKEN
* use certs from parameters
* use docker workspace & options
* add instance parameter
* implement branchName
* implement branchName
* remove duplicate default
* update docs
* fix TODOs
* remove merge mess
* fix code climate issue
* address comments
* respect custom defaults and custom config file name
* fix typo
* remove obsolete test cases
* disable CommonStepChecks
* check step config instead of context config
* remove TODOs
* respect jenkins PR envvars
* use value from stepConfig
* Update vars/sonarExecuteScan.groovy
* rename options to config
* correct type for options
* add test cases
* log sonar.options in debug message
Co-authored-by: Oliver Nocon <33484802+OliverNocon@users.noreply.github.com>
* iterate over found pom.xml files for maven projects, deploy found artifacts
* classifiers are found by testing prefix/suffix of artifact
* Remove additionalClassifiers option
* Improve documentation
* Fix fallback for finalBuildName
* Ignore mvn modules with packaging != pom but no target dir
* Fail when main artifact is missing, add more tests
Co-authored-by: Stephan Aßmus <stephan.assmus@sap.com>
* Fix typos
* Support aliases also for secrets
* Adapt & extend Unit Tests
* Output deprecation warning for param/secret aliases
... if the alias is marked as 'deprecated'.
* Config and Defaults via http(s)
Add the possibility to use http(s) urls as path to
* config.yml
* defaults
* Remove comment
* Addressing PR feedback
* Fix error handling
* Replace the default maximum request deadline with a default timeout on the transport level.
* Keep the possibility to set a maximum request deadline.
Artifacts to upload are assembled for MTA projects and Maven projects with optional application sub-module. Then maven deploy:deploy-file is used as backend to upload bundles of artifacts plus sub-artifacts.
Co-authored-by: Florian Wilhelm <florian.wilhelm02@sap.com>
* Golang step metadata: Config aliases for steps
This will ease following scenarios:
* config migration due to step name changes
* re-use of more general config, e.g. `mavenExecute` in `mavenBuild`
* fix CodeClimate finding
* Fix panic if original stage config does not exist yet
* add static code checks for maven based projects as:
* pmd plugin
* spotBugs plugin
* test modules as unit-tests and integration-tests will be ignored by default. Additional modules to ignore are configurable
* for pmd: rulesets and excludes are configurable
* for spotBugs: includeFilter and excludeFilter are configurable
ExecMockRunner and ShellMockRunner both needs an environment. "Extending"
here leads to "subclasses" for both cases. That is more long-winded since
it could be.
I don't understand why there should be a two dimensional array.
When dealing with envs we have normally a list containing entries like
[]string{"DEBUG=true", "HOME=/home/me"}
Having two dimensional env arrays would mean to have several alternate
environment in the tests at the same time. Don't think there is a need
for that.
* same behaviour for shellRunner and execRunner wrt errors and stdout
* replace shouldFail with shouldFailOnCommand
* [formatting only] format struct
* Move to regex for execptions and stdout
* shrink code
* Protecode as go implementation
Co-authored-by: Sven Merk <33895725+nevskrem@users.noreply.github.com>
Co-authored-by: Oliver Nocon <33484802+OliverNocon@users.noreply.github.com>
* Support yml and yaml extension of config files
* Update cmd/piper.go
Co-Authored-By: Christopher Fenner <26137398+CCFenner@users.noreply.github.com>
Co-authored-by: Christopher Fenner <26137398+CCFenner@users.noreply.github.com>
* Enhance telemetry reporting
* Use central telemetry data object
* Add duration
Co-Authored-By: Christopher Fenner <26137398+CCFenner@users.noreply.github.com>
* Use commonPipelineEnvironment in go binary
* Update groovy part incl. tests
* Rework structure and naming
* Support influx resources in steps
* Update tests and some cleanups
* Add correct defer handling
* Address PR feedback
* Fix test
* Update resources.go
Co-authored-by: Sven Merk <33895725+nevskrem@users.noreply.github.com>
* Remove redundancy: no duplicate stepName
Up to now: provided explicity via command line argument and also contained in the
step metadata (which are as a file also provided as command line argument).
Now it is retrieved from the metadata file.
* Step name is not a required parameter anymore
Since the step name is derived from the config file
* Adjust method signature in test
* Read the project config only if it exists
This avoid trying reading the file and have the control flow based
on errors. Beside that it helps troubleshooting when we have some
logging (debug level only).
* formatting only
* Adjust log level
There was some command parsing with failure in case it started with fail. That is
IMO less transparent. Now we prepare more explicit with a failure from outside. This
enables us to prepare an error like we expect it in the free wild.
* Export general configuration - part 2
First part in #956 missed to export the elements of the struct ...
* Add comment for exported struct
* Add function for opening config files
