Hi Dave,
Our patch touches code also changed by the patches that were recently committed.
That's likely what's causing this issue. We've rebased on top of the new state of master.
We had initially kept the yarn.lock .gitignored, but ran into an issue rather early on where an upgraded dependency introduced a regression.
Checking-in the yarn.lock provides the "know your dependency version" benefit of vendorizing code without vendorization's drawback of having to manually manage your dependencies.
It is safe to delete a yarn.lock before applying a patch, as you are authoring master. It provides a history of the versions of each dependency that were working at the point in time of the commit. By contrast, package.json provides approximate versions and leaves room for fixes/improvements by the dependency authors to be pulled in as they become available.
To run linter and tests:
The tasks that Grunt used to manage are now defined as a set of scripts in the package.json
After applying the patches--which may require deleting yarn.lock for the first patch--you should cd web && yarn install
Then yarn test will run the linter, jasmine specs, and python tests including feature tests, in that order, exiting early if there are failures/errors.
At the moment, the CheckForViewData test is failing on master as well as in each of these patches; that should be resolved as RM2477.
Thanks!
George and Matt