You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I think I can do it! Now that I have the command runner wired up, my tasks are:
this issue
figure out how docker_image should be set on Process in Python -- confirm it should be set for every single Process.
add environment: str field to targets
Then modeling should be done, other than Chris's rework of using subsystems for defaults. The rest is getting things to actually work properly, especially #16754.
I'm only working for 2 days next week, so we'll see how far I get.
Closes#16770.
Now, we dynamically determine what environment target to use at the root of the graph, and inject the value. This will always be the `__local__` target, or `None` if no targets are defined. That is, it cannot be a `docker_environment` etc. Then, callers can change the value to be a different environment by using `Get()` to inject a different `EnvironmentName` `Param`.
To avoid a rule graph cycle, this adds `WrappedTargetForBootstrap`.
[ci skip-rust]
Post #16721, the
EnvironmentName
is a parameter throughout the graph.ResolvedEnvironmentAlias
should now take the name/location ofEnvironmentName
.To accomplish that, we will need to pre-compute an
EnvironmentName
as a bootstrap step, to inject when running@goal_rule
s, etc.The text was updated successfully, but these errors were encountered: