Stu Hood
Stu Hood
Because `Pants` is bootstrapped in a platform dependent way (the rust code is always built on localhost), we will not be able to cross-build our integration tests across platforms anytime...
Add support for cross-building Python applications to `test`, `package` inside-of/for a Docker image
A few folks from Toolchain will be starting on this in the next few days. Here is a preliminary design doc: [Native Container Support](https://docs.google.com/document/d/1Nfjw153IR8iq9ltwQZfXxdSlMhp4MB_uG1grxeIxIaw/edit#), but there will be independent design...
Add support for cross-building Python applications to `test`, `package` inside-of/for a Docker image
A status update on this! As suggested by the linked issues/pulls above: 1. the rule graph changes necessary to support using multiple environments in a single build have landed (#16668,...
Add support for cross-building Python applications to `test`, `package` inside-of/for a Docker image
Note: I've removed `run` from the title here to capture the fact that a first version will not support "interactive" processes (`run`, `test --debug`, `repl`, etc), but will support standard...
Add support for cross-building Python applications to `test`, `package` inside-of/for a Docker image
Another update here. This week: * `docker` containers are now cached and reused * A preliminary implementation of making options environment-specific has landed. * `python_tests` targets gained an `_environment` field...
Add support for cross-building Python applications to `test`, `package` inside-of/for a Docker image
With #17129 and #17096 landing ~tomorrow, this issue will be completed, and ready for ([preview](https://github.com/pantsbuild/pants/issues/17355)) usage in the `2.15.x` release. See #17096 for the documentation for the Environments feature, which...
Mm... sorry for the trouble. This is because our [local process sandbox](https://github.com/pantsbuild/pants/blob/db07a12dd55dd5d9bda5518ac91961720582f7c6/src/rust/engine/process_execution/src/local.rs#L334-L341) doesn't use graceful shutdown for processes, and instead kills them immediately with SIGKILL. We do actually have an...
> So it sounds like we should use that code with local process sandbox cleanup, too? Yea.
>> I'm fine if we solve that (and other edge cases) in subsequent PRs. Maybe add an expected-to-fail test for the use case and open an issue? > > I...
@Eric-Arellano: The description looks great, thanks! @cosmicexplorer: If the existing rules might be helpful now, would you mind commenting on #7721 with a minimal set of things that @Eric-Arellano can...