Skip to content

Tags

Tags give the ability to mark specific points in history as being important
  • 0.22.0
    ### Changed
    
    * Allowed to call `initialize_needs()` and `initialize_dependencies()` on Sequence with empty parameter list,
      such that jobs explicitly does get an empty needs and dependency list.
  • 0.21.0
    ### Added
    
    * Added the `dependencies` keyword to Job, JobFilter and Sequence.
    
    ### Fixed
    
    * Fixed `set_artifacts` method on `Job` class.
  • 0.20.0
    ### Added
    
    * Added `allow_failure` feature on Jobs, JobFilter and Sequences.
    * Added method `set_needs` to Job, to allow setting empty needs array, as well as methods to initialize
      and override needs on Sequence level.
    
    ### Changed
    
    * Do `apk upgrade`  in Dockerfile - security over functionality.
  • 0.19.2
    ### Changed
    
    * Added need-dependencies for jobs in `gcip.addons.container.sequences.copy.CopyContainer`. This would improve speed in large pipelines
  • 0.19.1
    ### Fixed
    
    * Fixed typing of `Sequence.add_needs()` which now allows safely add sequences as needs to sequences.
  • 0.19.0
    ### Added
    
    * Documented the use of [prefilled variables](https://docs.gitlab.com/ee/ci/pipelines/index.html#prefill-variables-in-manual-pipelines) in combination with gcip in the user documentation.
    
    ### Changed
    
    * **BREAKING** The init method of the Rule class does not accept (unused) positional arguments anymore.
    * **BREAKING** Replaced the `gcip.core.Image` methods `set_tag()` and `set_entrypoint()` by the methods
      `with_tag()` and `with_entrypoint()`. The new methods return an altered copy of the original Image object.
      This is because the massive confusion when using centrally provided images
      (like `gcip.addons.container.PredefinedImages`) and accidentially modifying the object for the whole
      pipeline instead for a single place
  • 0.18.0
    ### Changed
    
    * Changed PredefinedImages to use tag arguments.
    
    ### Added
    
    * Added tag argument to class Image.
  • 0.17.0
    ### Added
    
    * Most of the `gcip.Job` attributes could be set in the `__init__()`-Method.
    
    ### Fixed
    
    * Fixed needs name generation bug where job names for 'needs' were wrongly generated. See #de9b3984 for further information.
  • 0.16.0
    ### Changed
    
    * Added more options to `mypy` job.
    
    ### Fixed
    
    * Fixed `mypy` job, `mypy` in versions newer `0.812` will output, that no typing stubs are installed. This leads to a failed pipeline job if there are external dependencies. Therefore the `mypy` job checks if `mypy` is installed if not it will be installed with a pinned version.
    * **BREAKING** Converted cdk jobs and sequences from functions to classes.
    * **BREAKING** Converted python jobs and sequences from functions to classes.
    * **BREAKING** Converted gitlab jobs from functions to classes.
    * Moved `scripts` to own module.
  • 0.15.3
    Fixed
    
    Missing mkdir -p $TAR_PATH in crane pull job.
  • 0.15.2
    ### Fixed
    
    * Using CI_COMMIT_REF_SLUG environment variable in crane, kankio to create the tag if none is given doesn't allow e.g. doted tages like `3.13` the SLUG will change the `.` to `-` so that the tag looks like `3-13`. With this fix the tag will be exactly be named after the Branch/Ref name.
  • 0.15.1
    ### Fixed
    
    * Missing `cache` property on `gcip.addons.container.sequences.*`.
  • 0.15.0
    ### Added
    
    * Added contextmanager methods to Pipeline class.
    
    ### Changed
    
    * Changed container addons from simple functions to more enhanced objects. For a comprehensive explanation about this change see MR !113.
  • 0.14.1
    ### Fixed
    
    * In accordance to MR-105, there was the needs name generation missing. The needs job name was generated with the old style. Job name after Stage, now it is Job name before stage. Most relevant text comes first for needs too.
  • 0.14.0
    ### Changed
    
    * **BREAKING** Removed obsolete `gcip.tools.evaluate_git_tag_pep404_conformity.py` file.
    * #43 - `gcip.addons.container.DockerClientConfig.add_raw()` does now deep merge given raw configuration dict.
    * **BREAKING**  Made `gcip.core.job.Job.copy()` private, as it is only intended for internal use.
    
    ### Fixed
    
    * DockerClientConfig used in `gcip.addons.container.seqences.*` was modified after passing object to function, that leads to wrong path's after first modification in `kaniko` job.
  • 0.13.0
    ### Added
    
    * #86 - Added possibility to reference `pipeline`s in `Need`s.
    * Added `Sequence.find_job()` to find all jobs in a Sequence (or Pipeline) with matching attributes. Returned
      jobs could then be modified.
    
    ### Changed
    
    * Changed order of job name components. Most relevant text comes first. See discussion in [MR-105](https://gitlab.com/dbsystel/gitlab-ci-python-library/-/merge_requests/105).
  • 0.12.1
    ### Fix
    
    * #85 - using `CI_PROJECT_DIR` as artifact path only, raises an error
  • 0.12.0
    ### Added
    
    * Added options to the CDK deploy job and sequence making `--strict` optional.
  • 0.11.0
    ### Added
    
    * Added auto download for `gcip.addons.security.sops.sops_export_decrypted_values`.
    * Added new container sequence to allow copying a container image between registries.
    
    ### CHANGED
    
    * Removed `name` argument from `gcip.addons.container.sequence.build.full_container_sequence` jobs.
  • 0.10.0
    [0.10.0] - 2021-05-04
    
    Added
    
    Added all public available classes to __all__ variable.
    
    
    Fix
    
    Added boto3~=1.17 to setup.py as a dependency for addons.
    Added possibility to use self exported environment variables if not in pipeline environment.