You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
If files are changed in any sciebo resources that are used for testing, pipelines do not notice until next run. If changes in sciebo are done accidentely, the CI won't be able to get the source of the failure.
We should think about reducing the file sizes that are used for testing and move all the files back to github to ensure pipelines are triggered once files have changed. Also, sciebo may be limited in restoring data.
The text was updated successfully, but these errors were encountered:
I would strongly disagree to put the resource files for testing in the bim2sim repository (main repository) for multiple reasons, instead @Cudok and I developed the following approach:
create a second github repository for all test resources (and mirror this to gitlab)
create a pipeline there that runs on merge request and clones the current development (and main? but no other branches for simplification) branch and tests the code with the changed test resources
If the pipeline fails not merge is possible -> we will always know when something goes wrong and we will have history of the test resources. A pipeline will look like this:
test_Base:
image: $CI_REGISTRY_BIM2SIM/environment:development # this variable needs to be creates as the docker images are not stored in the test resources repository
stage: test
script:
- clone bim2sim repo /bim2sim-coding/ # clone bim2sim coding repository from main repo
- mv ./* /bim2sim-coding/test/resources # copy the changed test resources into the bim2sim-coding
- coverage run -m unittest discover /bim2sim-coding/test # run tests with changed test resources
- coverage report -i # give report
If files are changed in any sciebo resources that are used for testing, pipelines do not notice until next run. If changes in sciebo are done accidentely, the CI won't be able to get the source of the failure.
We should think about reducing the file sizes that are used for testing and move all the files back to github to ensure pipelines are triggered once files have changed. Also, sciebo may be limited in restoring data.
The text was updated successfully, but these errors were encountered: