This is the future source of truth for building the OpenShift Pipelines operator with Red Hat Konflux CI instance.
This repository vendors tektoncd/operator
and gets image references from other "vendored" project from openshift-pipelines.
The list of projects that "contribute" to this are:
openshift-pipelines/tektoncd-pipeline
(upstream:tektoncd/pipeline
openshift-pipelines/tektoncd-triggers
(upstream:tektoncd/triggers
openshift-pipelines/tektoncd-chains
(upstream:tektoncd/chains
openshift-pipelines/tektoncd-results
(upstream:tektoncd/results
openshift-pipelines/tektoncd-cli
(upstream:tektoncd/cli
openshift-pipelines/tektoncd-hub
(upstream:tektoncd/hub
openshift-pipelines/manual-approval-gate
openshift-pipelines/pipelines-as-code
openshift-pipelines/tekton-caches
openshift-pipelines/opc
In addition, some of the configuration of this repository are generated from the hack
project.
project.yaml
holds the source of truth of most things that are generated. It stores all the component's built images (not released yet) as well as version numbers, …head
holds the commit SHA fromtektoncd/operator
that is currently vendored.upstream
holds the unchanged vendored code fromtektoncd/operator
openshift
holds openshift/openshift-pipelines specific code such asdockerfiles
for the operator's images.bundle
for the bundle image. A lot of files in this repository are generated from the upstream code and somehack
scripts.index
for the index image.rpms
for the "prefetch" of rpms (as we build everything in an hermetic way)
.github
holds GitHub workflow configurations. Some of them are automatically generated by thehack
project..konflux
and.tekton
hold konflux and tekton manifests, and are mostly coming from thehack
project.
TBD
- We need to be able to re-build everything from scratch
- the operator
- the operands, such as pipeline, …
- We need to be able to push them, generate the bundle and then the image index
- We need to have this all happen in konflux…
- … that doesn't support
ko
today
- … that doesn't support
- We need to track upstream release and map to it ours.
- We need to track release branches as well