Currently, we send all files as a list but it can be problematic with
files that can't be evaluated properly. Instead of crashing the entire
extraction process, we will send a file at a time for eval so we can
just bypass files causing issues.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
Since we're checking out the `pull_request_target` and not the
`pull_request` branch. We can't evaluate the files from the PR branch.
We can skip the files for evaluation without crashing the entire
extraction.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
Break the workflow into multiple scripts to make it easier to test /
maintain. Also fix the remove reviewer process to not review reviews
from people that were manually requested.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
Simplify extraction and generation of file. We dont need the comments
and can leverage the appropriate lib.generator function.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
Should remove maintainers that existed, but weren't assigned to modules.
Will add maintainers that we missed during last refactoring.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
Previously, we had to hack together some string matching to identify and
retrieve the maintainers in the repo. We can just eval the modules to
retrieve the list of maintainers more accurately.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
Used to eval home-manager and fetch meta.maintainers to accurately
retrieve maintainers for the master list.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
Create a script for generating the master maintainer list we will use
for inviting / pinging on changed files.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
Using ./. forces Nix to copy the flake to the store a second time.
Using bulitins.path like this, forcing the name to be "source", reduces the extraneous copy.
In most cases where this function is used, suppressing only the standard
output is more appropriate. Culling diagnostic output hides error
messages and makes debugging more difficult and confusing.
`$DRY_RUN_NULL`, which the `--silence` flag replaced, was used both for
suppressing standard output on its own, and for doing so along with
diagnostic output; however, when the `run` function was added this
distinction was lost, and both outputs would be discarded.
This reintroduces the needed functionality, and changes usages of
`--silence` to `--quiet` where previously only standard output was
suppressed, or where this should have probably been the case anyway.
Change-Id: Ifb1b52a1d1eea0117261c782d686ad7c71b43162
Specifically, if `tput colors` fails with an error we treat that as if
the terminal does not support colors.
Fixes#423
Suggested-by: PhotonQuantum <self@lightquantum.me>
