I'm beginning to think that the change to the CI trigger in #49 is causing the problem.
We're seeing that in almost all cases, the CI job is running against master not the latest commit from the branch.
You can see this in that only one of the recent workflow runs, it is reporting a branch name. This is the only test run that successfully cat the time entries file.
So, I'm going to put back BOTH events and we should expect that they'll pass the linting because they pull in the latest code but then the tests should fail.
I think in the long run, I need to do development not from my fork, but directly from this repo now that I'm a collaborator.
I need to step away for a few hours so I'm going to let this sit. It will also let me think on it for a while too.
I'm beginning to think that the change to the CI trigger in #49 is causing the problem.
We're seeing that in almost all cases, the CI job is running against
master
not the latest commit from the branch.You can see this in that only one of the recent workflow runs, it is reporting a branch name. This is the only test run that successfully
cat
the time entries file.After some more reading, I'm not sure that PRs from forked repositories are allowed to access the secret toggl API token even with
pull_request_target
. See the announcement when the feature was launched. https://github.blog/2020-08-03-github-actions-improvements-for-fork-and-pull-request-workflows/#improvements-for-public-repository-forksSo, I'm going to put back BOTH events and we should expect that they'll pass the linting because they pull in the latest code but then the tests should fail.
I think in the long run, I need to do development not from my fork, but directly from this repo now that I'm a collaborator.
I need to step away for a few hours so I'm going to let this sit. It will also let me think on it for a while too.