Hi stranger! We've written this document for:
- Active maintainers of Mocha
- Prospective maintainers of Mocha
- Anyone curious about how Mocha's maintainers maintain Mocha
The purpose of this document is to describe our processes. We want to avoid conflicts and confusion around "unwritten rules". In our opinion, the most straightforward way to address this is to write them down. This also happens to be the most straightforward way to change them!
To assist in eliminating ambiguity, we will define some terms.
Anyone involved with Mocha will fall into one of these buckets: user, contributor, and maintainer.
A "user" for the purpose of this document is any individual developer who consumes Mocha to write and/or execute tests. A user interacts with contributors. A user interacts with the software, web site, documentation, etc., which these contributors provide.
As a user, you're expected to follow the code of conduct when interacting in Mocha's "official" social spaces. This includes:
- Any channel under the
mochajs
Discord - Any project under the
mochajs
organization on GitHub - Any future social, in-person, or online events which Mocha might organize
This is the most important thing:
You don't have to write code to be a contributor!
A "contributor" is any individual who has given back in some way to the project and its community. Contributions include (but are not limited to):
- Reporting bugs which follow the reporting guidelines
- Suggesting and debating enhancements that have wide applicability
- Helping others with Mocha-related questions on our Discord, StackOverflow, or other sites
- Sending pull requests which fix bugs, improve documentation, improve developer experience, improve code quality, and/or implement requested enhancements
- Reviewing code on pull requests
- Providing design assets
- Posting a tutorial on a personal blog or blogging site
- Suggesting usages for project funds
- Organizing a "Mocha-branded" event or workshop
- Recruiting more contributors! Don't spam.
- Researching the user base, getting feedback, etc. Don't spam.
A contributor is usually a user as well, but this isn't a hard-and-fast rule. A contributor is also expected to adhere to the code of conduct as a user would.
As you can see, it's wide open! Think of it another way: if you are adding value to Mocha, then you are a contributor.
Due to the nature of GitHub, it's a challenge to recognize those who've made contributions elsewhere on the web, or even contributions of the "non-code" variety. If you know of any great contributions which have gone unnoticed, please bring them to the maintainers' attention!
A donation is also a great way to help Mocha if you want to help sustain OSS, but can't find time to contribute in other ways, or just want to say "thanks!"
We love our backers and sponsors! 💕
A maintainer has certain "rights" (or "permissions") to the Mocha project and other projects under the mochajs
organization.
There's no way to dance around this: with these rights come increased responsibilities.
However, there is no expectation of a standard of technical ability to be a maintainer of Mocha. This doesn't imply a lack of technical oversight--every pull request will eventually be reviewed.
If you think you aren't experienced enough to maintain a project like Mocha, you are incorrect. The only requirements are the above responsibilities and a desire to help the project. It bears repeating:
You don't have to write code to be a maintainer!
Maintainer is synonymous with "Collaborator" and/or "Owner" in GitHub parlance.
As a maintainer, you are expected to not just "follow" the code of conduct, but embody its values. Your public behavior, whether in the physical or virtual world, reflects upon the project and other maintainers.
If you don't understand the code of conduct, or why it exists, it is your responsibility to educate yourself. This does not imply the CoC is immutable.
Furthermore, a maintainer is a contributor who contributes regularly, or expresses a desire to do so. That could be every day--but it might be once a week, or even once a month. Your boss doesn't work here; contribute as often as you wish. We are all people with Real Lives, and for many of us, contributing to OSS is just a hobby!
Finally, a maintainer must help define what makes Mocha "Mocha". At minimum, a maintainer must understand the current definition (if a maintainer is not interested in decision-making). Some of these questions include:
- What's the scope of Mocha?
- Where should we focus efforts?
- What's urgent, what can wait?
- What can we break? What's off-limits?
- What user feedback is valuable? What isn't?
As maintainers, we work together to learn about the nature of these questions. If we try hard enough, we even come to some answers!
A maintainer must also have 2FA enabled on their GitHub account.
If you think that you aren't familiar with mocha's internals enough to contribute, please watch This walkthrough video!
You may choose to do zero or more of these at their discretion:
- Merge pull requests
- Modify issues (closing, adding labels, assigning them other maintainers, etc.)
- Modify GitHub Projects
- Cancel builds, restart jobs, or otherwise interact with our CI server(s)
- CRUD operations on GitHub integrations
- Participate in the decision-making process
- Add new maintainers to the team
- Tag releases and publish Mocha to npm
While maintainers have the ability to commit directly to the
main
branch, this is to be avoided if any other maintainer could reasonably take issue with the change, or the change affects Mocha's API or output. For example, a spelling correction inCHANGELOG.md
may not require a pull request. A change to a reporter's output most certainly would! Maintainers are trusted to use their best judgement; if unsure, err on the side of caution.
Some maintainers will have full admin rights to the mochajs org and/or will have access to publish to npm.
- Those with publish access are expected to use npm's 2FA.
- This level of access will be granted by the current owners to those maintainers who have earned the project's trust.
Mocha follows a consensus-seeking decision-making process. In other words, all maintainers attempt to come to agreement. If that fails, we decide by a simple vote.
Active maintainers will make an effort to solicit feedback from others before making important or potentially controversial decisions. Given the varying geographical distribution and availability of the maintenance team, we resolve to do the best we can to solicit feedback.
In other words, to have your opinion heard, participate regularly. The rest of the team won't wait on feedback that isn't necessarily forthcoming!
Maintainers will mainly gather in the Mocha Discord. This is a public Discord, and anyone can join. Videoconference (or audio) calls may happen on a regular or irregular basis, as schedules allow. This is mainly because we have Real Lives and time zones suck.
All new issues will need to be triaged, and pull requests must be examined. Maintainers must understand Semantic Versioning ("SemVer"), as Mocha follows it strictly.
If you see an issue or PR that could use some labels, please add them!
The TL;DR of Semantic Versioning is:
- MAJOR version when you make incompatible API changes,
- MINOR version when you add functionality in a backwards-compatible manner, and
- PATCH version when you make backwards-compatible bug fixes.
Pull requests must have one of these three (3) labels:
semver-patch
for backwards-compatible bug fixes, documentation, or anything which does not affect a "production" (npm install mocha
) installation of Mochasemver-minor
for backwards-compatible new features or usability/interface enhancementssemver-major
for backwards-incompatible ("breaking") changes to the API
A PR which introduces a breaking change is considered to be semver-major
, regardless of whether it's a bug fix, feature, or whatever.
For the purposes of the above definitions, Mocha has some unique considerations, and includes the following in its definition of "API":
- Mocha's documented, programmatic interface which is not explicitly tagged with
@private
- Mocha's machine-readable reporter output
- Mocha's default settings
- Mocha's command-line options
- The environments which Mocha supports; this includes:
- Browser versions
- Node.js versions
- Compatibility with popular module loaders (e.g., AMD)
Err on the side of the user; breaking changes to private APIs will be semver-major
, if and only if they are known to be consumed by actively developed project(s).
Examples of a breaking changes might be:
- Throwing an
Error
where one wasn't thrown before - Removing a command-line option or alias
- Removing an environment from the CI configuration
- Changing the default reporter!
- Changing defaults in a way which would cause tests which were previously successful to start failing, or a failing test to start passing
- The exception is fixing likely false-positives
- A good example would be changing the default
timeout
value
Issues should be filed according to one of our GitHub Issue forms.
If any required information is missing, add the status: waiting for author
label and politely ask for the missing information.
For all issues, apply the following labels based on which area(s) the issue pertains to:
area: async
: Issues around Mocha's asynchronous usagearea: browser
: Issues unique to a browser environmentarea: parallel
: Issues around Mocha's parallel modearea: reporter
: Usually concerning Mocha's outputarea: repository tooling
: Issues around Mocha's CI, own test suite, or other internal toolingarea: security
: Involving vulnerabilities, actual or potentialarea: windows
: Windows-specific issues, particularly around path discrepancies
Additionally:
good first issue
: If the implementation is likely doable by someone who's never contributed to Mocha (or potentially any other open source project) beforestatus: duplicate
: If an equivalent issue was already filed, add this label, close as not planned, and comment with something likeduplicate of #<other-issue-number>
status: in discussion
: Add this whenever the issue is blocked on community input and/or deeper discussionsstatus: in triage
: Added on new issues; re-add this whenever the issue is awaiting maintainer attentionstatus: waiting for author
: Add this whenever the issue is blocked on something from the author
Bug reports should include a way to reproduce the issue that someone who is not deeply familiar with Mocha can work with locally.
Depending on that reproduction, remove status: in triage
and add the following label(s) in addition to the auto-added type: bug
:
- If the bug is valid and reproduction works: add
status: accepting prs
- If the bug might be valid but the reproduction isn't workable:
- Add
status: waiting for author
- Politely comment explaining why the reproduction isn't workable
- Add
- If the bug might be valid but it's not clear whether it's worth it:
- Add
status: in discussion
- Explain that it might be worth it and that more community input is needed
- Add
- If the bug is clearly not worth it or valid, explain why, close the issue as not planned, and:
- If it isn't a bug at all: add the
invalid
label - If it is roughly a bug but isn't something that can or should be fixed: add the
status: wontfix
label
- If it isn't a bug at all: add the
Documentation reports should clearly indicate a gap or problem that should be addressed in documentation.
Triage documentation issues similar to bugs and/or feature requests - documentation is its own form of product area.
Remove status: in triage
and keep the auto-added area: documentation
label.
Feature requests should include a compelling reason why we should spend the maintenance time on the feature. Given that Mocha is prioritizing stability over growth, this can be a high bar.
Depending on the reasoning, remove status: in triage
and add the following label(s) in addition to the auto-added type: feature
:
- If the reasoning is valid and seems worth the maintenance cost: add
status: accepting prs
- If the reasoning is unclear:
- Add
status: waiting for author
- Politely comment explaining what's missing
- Add
- If the feature might be valid but it's not clear whether it's worth it:
- Add
status: in discussion
- Explain that it might be worth it and that more community input is needed
- Add
- If the feature is not valid, explain why, close the issue as not planned, and:
- If it isn't a feature request at all: add the
invalid
label - If it is roughly a feature request but isn't something that can or should be implemented: add the
status: wontfix
label
- If it isn't a feature request at all: add the
Issues filed about improvements to Mocha's internal development processes.
These can be more informally discussed by maintainers.
Remove status: in triage
and keep the auto-added area: repository tooling
.
Our issue tracker is not the right place to ask questions.
If an issue is filed that seems like it's more of a question, remove status: in triage
, add the type: question
label, politely direct the user to the ❓ Got a Question? section, and close the issue as not planned.
If it's not a Mocha problem (people tend not to believe this), you may want to show a counter-example. It's often helpful to direct the issue author to the responsible project, if you can determine what that is.
If this issue seems to be repeatedly asked, add the faq
label.
This may also apply to questions which receive a lot of 👍 reactions.
Write "closes #" or "resolves #" in a commit or PR to have the original issue closed automatically once the PR is merged.
For any issue which is a duplicate, write "duplicate of #" in a new comment, and close the issue. Read more about marking issues as duplicates.
If the issue is a support question, and you believe it has been answered, close the issue.
If the issue is not Mocha-related, and/or a bug cannot be confirmed, label it invalid
and close.
All maintainers should be courteous and kind. Thank the external contributor for the pull request, even if it is not merged. If the pull request has been opened (and subsequently closed) without discussion in a corresponding issue, let them know that by creating an issue first, they could have saved wasted effort. Clearly and objectively explain the reasoning for rejecting any PR.
If you need more information in an issue, nicely ask the user to provide it. Remind them to use the issue/PR templates if they have not.
Use GitHub's code review features. Requesting a review from another maintainer may or may not actually result in a review; don't wait on it. If the PR cannot move forward without input from a certain maintainer, assign them to the PR.
There will be jerks.
These are users who feel the Mocha project and its maintainers owe them time or support. This is incorrect.
However, this behavior is often indicative of someone who is "new" to open source. Many just don't know better. It is not your responsibility to educate them (again, you owe them nothing).
Here are some suggestions:
- If u mad, wait 20 minutes before writing a comment.
- "Kill them with kindness". Explain how they are presenting themselves; maybe link to a good article or two about it.
- Don't make it about "users vs. maintainers". Treat them like a potential future maintainer.
- Avoid adding to the drama. You could try to reach out privately; email may be in their GitHub profile. You will likely never hear from that individual again (problem solved)
- If an issue is getting out of control, lock it.
- If someone is repeatedly rude and does not correct their mistakes, you may ban them from participating in the
mochajs
org. If you do not have permission to do so, contact one which does (an "owner").
This section is theoretical, as it's yet to happen.
- Inform the individual of the violation; link to the CoC
- Follow up with OpenJS Foundation for further guidance
- Repeated violators will be banned inasmuch as that is technically possible
- No maintainer nor contributor is exempt from the CoC
main
is the only maintained branch in mochajs/mocha
or any of the other repos.
main
is the only branch to which force-pushing is disallowed.
Maintainers may push new branches to a repo, as long as they remove them when finished (merging a PR will prompt to do so).
Please please please delete old or unused branches.
We prefer to squash merge PRs. Requiring users to keep clean histories for rebasing would be a large ask that we don't feel justifies the benefits.
Upon acceptance of a PR, you must assign it a milestone.
If you know that the PR is breaking, assign it to a new or existing milestone correlating with the next major release.
For example, if Mocha's current version is v6.5.2, then this milestone would be named v7.0.0
.
Likewise, if the PR is semver-minor
, create or use a new milestone correlating to the next minor release, e.g., v6.6.0
.
If it's unclear what the next milestone will be, use or create a milestone named next
.
This milestone will be renamed to the new version at release time.
By using milestones, we can cherry-pick non-breaking changes into minor or patch releases, and keep main
as the latest version.
This is subject to change, hopefully.
It's easier to release often.
- Decide whether this is a
patch
,minor
, ormajor
release. - Checkout
main
in your working copy & pull. - Modify
CHANGELOG.md
; follow the existing conventions in that file. Use the "pull request" number, unless there isn't one. You do not need to add Markdown links; this is done automatically.- You can omit stuff from
CHANGELOG.md
that was done by a maintainer, but would have no interest to consumers of Mocha. - If the changes aren't of interest to consumers but were not made by a maintainer, reference them anyway. It's cool to give attribution!
- You can omit stuff from
- Use
npm version
(usenpm@8+
) to bump the version; seenpm version --help
for more info. (Hint--use-m
: e.g.,npm version patch -m 'Release v%s'
)- This command will update the list of authors (from the Git history) in
AUTHORS
, and add GitHub links toCHANGELOG.md
. - These changes are then added to the Git "stage" and will be added to the commit.
- This command will update the list of authors (from the Git history) in
- Push
main
toorigin
with your new tag; e.g.git push origin main --tags
- Copy & paste the
CHANGELOG.md
lines to a new GitHub "release". Save release as draft. - Meanwhile, you can check the build on Travis-CI and GitHub Actions.
- Once the build is green, you'll want to trigger an update of
mochajs.org
: - If you're doing a prerelease, fast-forward the
next
branch tomain
, and push it. This updates https://next.mochajs.org. That's all. - If this is NOT a prerelease, fast-forward the
mochajs.org
branch tomain
and push it. This updates https://mochajs.org. - If this is a "final" release (the first release of a major after one or more prereleases) then remove the
next
tag from npm vianpm dist-tag rm next
.
- Once the build is green, you'll want to trigger an update of
- Finally, you're satisfied with the release notes, open your draft release on GitHub, then click "publish."
- Back in your working copy, run
npm publish
. If you're doing a prerelease, ensure that you use--tag=next
. - Announce the update on Twitter or just tell your dog or something.
Note: there are too many steps above.
The OpenJS Foundation retains copyright of all projects underneath the mochajs org. The Foundation does not influence technical decisions nor the project roadmap. It is, however, charged with ensuring the continued vitality and sustainability of projects under its banner.
As a maintainer, you have access to the resources the OpenJS Foundation provides.
Mocha collects donations via OpenCollective. As a maintainer, you may help decide how the funds are used. These decisions are made via a consensus-seeking process, much like any other decision.
Expense transparency is built in to OpenCollective.
Questions? Ask in the Mocha Discord!