fix(commitlint): skip task on master

[chrisbottin]

The rule between master and branch should be different to take the PR reference in consideration when merging.

---

This PR should be fixing the current build failure on master due to the latest commit message fix: report launcher process error when exit event is not emitted (#3647) being more than 72 characters (https://travis-ci.org/github/karma-runner/karma/jobs/757457475).

---

Problem

The problem is that the same header-max-length rule is applied to master and the PR branches and when a PR branch is merged the commit message has the PR reference appended to it making the length longer and potentially breaking the commitlint on master

Proposed Solution

Using different header-max-length rules for master and branches.

• header-max-length rule on master set to 80 characters to temporary fix the build on master and then submit another PR to set the header-max-length rule on master set to 72 characters.
• header-max-length rule on branches set to 64 characters (72 - 8 characters for the PR reference)

[devoto13]

Hm, maybe it is better to simply disable commit message validation on master? 🤔

[AppVeyorBot]

✅ Build karma 2926 completed (commit c076f6b558 by @chrisbottin)

[karmarunnerbot]

✅ Build karma 529 completed (commit c076f6b558 by @chrisbottin)

[karmarunnerbot]

✅ Build karma 528 completed (commit c076f6b558 by @chrisbottin)

[chrisbottin]

Hm, maybe it is better to simply disable commit message validation on master? 🤔

@devoto13 yes we could do that but do you mean permanently? If we are happy with having no commitlint running on master then we can just disable it (I can submit another PR) ... otherwise we do need this PR to have different header-max-length rules.

Let me know

[devoto13]

Yes, permanently. If you think about it, the check is not really actionable: one can't revert or fix the commit in master branch. So we can only run it in PRs/branches where we can prevent a bad commit message. But it's just my opinion. It's up to @johnjbarton to decide which approach to use.

[johnjbarton]

Yes @devoto13 is right, I like the no-master solution better. @chrisbottin can you do it?

[karmarunnerbot]

✅ Build karma 535 completed (commit 138b283a52 by @chrisbottin)

[AppVeyorBot]

✅ Build karma 2932 completed (commit 138b283a52 by @chrisbottin)

[karmarunnerbot]

✅ Build karma 534 completed (commit 138b283a52 by @chrisbottin)

[AppVeyorBot]

✅ Build karma 2933 completed (commit 339ac71f89 by @chrisbottin)

[karmarunnerbot]

✅ Build karma 536 completed (commit 339ac71f89 by @chrisbottin)

[karmarunnerbot]

✅ Build karma 535 completed (commit 339ac71f89 by @chrisbottin)

[chrisbottin]

@johnjbarton done as requested.

[johnjbarton]

Thanks, sorry for the delay lost track of this one.

[karmarunnerbot]

🎉 This PR is included in version 6.1.2 🎉

The release is available on:

• npm package (@latest dist-tag)
• GitHub release

Your semantic-release bot 📦🚀