Coq: How to get a summary of why CI fails?

[herbelin]

This is to continue a discussion started at PR#876.

When seeing a red cross signalling a failure of Travis or Appveyor in the list of PRs, it is tedious to know what the cause of the error, and frustrating when this error happens to often be an error not related to the PR itself, such as timeout, opam failure, missing shell script, network failure, ...

Travis and gitlab send mails for tests made on personal branches (with the gitlab mails incidentally more detailed as shown by @SkySkimmer).

Can these mails e.g. be used to report the status on the main PR page? Or at least to report a more precise status in the "check" subwindow at the bottom of individual PR pages?

[mattam82]

The gitlab mail seems like it has everything we need.

[Zimmi48]

Incidentally, this is linked to https://github.com/travis-ci/travis-ci/issues/8171, https://github.com/travis-ci/travis-ci/issues/8136 and https://github.com/travis-ci/travis-ci/issues/5035.

[mattam82]

https://github.com/wizeline/travieso seems to do the detailed reporting

[Zimmi48]

I like very much the detailed reporting by CircleCI. If we decide after testing it for some time that it is a good fit for us, this means that we can close the issue without installing any external service like Travieso.

[Zimmi48]

Using CircleCI had solved this for us. We are planning to change again the main CI system to go to GitLab CI though and this will have the same issue that we had with Travis. I don't expect anything like Travieso to exist for GitLab CI (given that even the native support connecting GitLab CI to GitHub is brand new). But we could open a feature-request on GitLab...

[ejgallego]

Thanks the work of many people this issue is now solved.