Conversation
| return " - ".join(parts) | ||
|
|
||
|
|
||
| def step_key(prefix, env, index=None): |
There was a problem hiding this comment.
I'm a bit concerned here that there are possible duplicate step names depending on the configuration of the matrix. As I've run in to before, if you try to dynamically build steps and they get the same name BK will reject the pipeline when you try to load it. I'm assuming that is wha the index key is for? Essentially i'm wondering how we guarantee that we will have unique step names, especially in the case where multiple cells have the same versions but different ENV vars.
| # Jobs format: explicit env strings per step | ||
| for i, group_cfg in enumerate(base_matrix["jobs"]): | ||
| groups.append(build_base_jobs_group(group_cfg, default_env, i)) | ||
| else: |
There was a problem hiding this comment.
How is this whole block ever encountered? I'm not seeing how jobs wont be in base_matrix. I must be minsunderstadning what base_matrix is...
There was a problem hiding this comment.
I have the same question. How does a plugin run into "stream" format? Are you expecting each plugin has a local base-pr-test-matrix.yml to override the share one? But looking at the command SHARED_CI_DIR=/tmp/ci PLUGIN_DIR=$(pwd) I don't see how the merge happens.
|
@mashhurs Could you include sample input files, commands, and output results to demonstrate how "stream" and "job" formats are triggered? |
3 participants
Release notes
What does this PR do?
Why is it important/What is the impact to the user?
Checklist
Author's Checklist
How to test this PR locally
Related issues
Use cases
Screenshots
Logs