A Tekton Enhancement Proposal (TEP) is a way to propose, communicate and coordinate on new efforts for the Tekton project. You can read the full details of the project in TEP-1.
A standardized development process for Tekton is proposed in order to
- provide a common structure and clear checkpoints for proposing changes to Tekton
- ensure that the motivation for a change is clear
- allow for the enumeration stability milestones and stability graduation criteria
- persist project information in a Version Control System (VCS) for future Tekton users and contributors
- support the creation of high value user facing information such
as:
- an overall project development roadmap
- motivation for impactful user facing changes
- ensure community participants are successfully able to drive changes to completion across one or more releases while stakeholders are adequately represented throughout the process
This process is supported by a unit of work called a Tekton Enhancement Proposal (TEP). A TEP attempts to combine aspects of the following:
- feature, and effort tracking document
- a product requirements document
- design document
into one file which is created incrementally in collaboration with one or more Working Groups (WGs).
This process does not block authors from doing early design docs using any means. It does not block authors from sharing those design docs with the community (during Working groups, on Slack, GitHub, ….
This process acts as a requirement when a design docs is ready to be
implemented or integrated in the tektoncd projects. In other words,
a change that impact other tektoncd projects or users cannot be
merged if there is no TEP associated with it.
This TEP process is related to
- the generation of an architectural roadmap
- the fact that the what constitutes a feature is still undefined
- issue management
- the difference between an accepted design and a proposal
- the organization of design proposals
This proposal attempts to place these concerns within a general framework.
See TEP-1 for more details.
The TEP OWNERS are the main owners of the following projects:
To create a new TEP, use the teps script:
$ ./teps/tools/teps.py new --title "The title of the TEP" --author nick1 --author nick2The script will allocate a new valid TEP number, set the status to "proposed" and set the start and last updated dates.
Note that the picked number is not "locked" until a PR is created.
The PR title shall be in the format TEP-XXXX: <tep-title>.
To help a TEP to be approved quickly, it can be effective for the initial content of the TEP to include the high level description and use cases, but no design, so reviewers can agree that the problems described make sense to address before deciding how to address them. Sometimes this can be too abstract and it can help ground the discussion to include potential designs as well - but usually this will mean people will want to agree to the design before merging and it can take longer to get consensus about the design to pursue.
The TEP PR might fail CI if a TEP number conflict is detected, or if
there is a merge conflict in the TEP table. In case that happens, use
the teps.py renumber command to refresh your PR:
./teps.py renumber --update-table -f <path-to-tep-file>
The command will update the TEP in the file name and content with a new available TEP number, it will refresh the TEPs table and it will present a list of git commands that can be used to update the commit.
TEP should be merged as soon as possible in the proposed state. As
soon as a general consensus is reached that the TEP, as described
make sense to pursue, the TEP can be merged. The authors can then add the
design and update the missing part in follow-up pull requests which moves the TEP to implementable.
Reviewers should use /approve to indicate that they approve
of the PR being merged.
TEP must be approved by at least two owners from different companies. Owners are people who are maintainers for the community repo. This should prevent a company from force pushing a TEP (and thus a feature) in the tektoncd projects.
Whenever possible, a TEP should be approved by all assignees to the PR. We may fall back on the strict requirement of approvals from at least two different companies in the case of an unresponsive assignee. TEP PRs shouldn't be merged if an assignee has strong objections.
TEP PRs that don't affect the proposed design (such as fixing typos, table of contents, adding reference links, or marking the TEP as implemented) do not need to meet these approval requirements. A single reviewer can feel free to approve and LGTM changes like these at any time. PRs marking a TEP as "implementable" should still meet the approval requirements, as this label signifies community agreement that the proposal should be implemented.
Tekton team members regularly review TEP PRs (those with the "tep" label) during the API working group.
- At this meeting, we try to find assignees to review TEP PRs, discuss any TEP PRs that need discussion, and merge any TEP PRs that have met the approval requirements.
- Reviewers assigned during the API working group should aim to give feedback by the next API working group meeting.
The TEP author can find reviewers as soon as the PR is created. Some great ways to find reviewers and publicize your proposed changes are:
- Reaching out to the stakeholders or any community maintainers through the PR comments
- The "tep" channel on slack
- The tekton-dev mailing list)
TEP collaborators are permitted to be reviewers.
Once all assigned reviewers have approved the PR, the PR author can reach out to one of the assigned reviewers
or another "reviewer" for the community repo to merge the PR.
The reviewer can merge the PR by adding a /lgtm label.
- If a contributor adds "lgtm" before all assignees have had the chance to review, add a "hold" to prevent the PR from being merged until then.
- Note: automation prevents the PR author from merging their own PR.
If the TEP has undergone substantial changes since any reviewers have approved it, the author should publicly confirm with the relevant reviewers that they are OK with these changes before the PR is merged.
Why don't we use GitHub reviewers instead of assignees? If we want to do that we need to turn off Prow's auto assignment of reviewers; there is no guarantee the auto assigned reviewers are the appropriate reviewers. See discussion.