Stu Hood

Results 228 comments of Stu Hood

> 1. Most `lint` partitioners look like [this](https://github.com/pantsbuild/pants/blob/cace8518a43646ac0e84c4f841cee01e46b1defc/src/python/pants/backend/shell/lint/shellcheck/rules.py#L43). I want to provide a generic "rule_maker" for that however the Subsystem param can't be known statically and is only known at...

> Bingo! Although it'd be a subset of `@union` members, as requested by plugin authors. It could be a separate `@union` probably: `LinterWhichUsesADefaultPartitioner`? Is that worth sketching out as an...

> See also use-case 2. Use case 2 also appears to be a `@union` use case, but maybe I'm misreading it. > I think this is much more general and...

@Kaushik-Iyer : Sorry for the delay! This would look a lot like the issue I linked above: https://github.com/pantsbuild/pants/pull/16174/files Essentially, when computing the `globals` to use when evaluating `BUILD` files, you...

Oof. Not terribly excited about that, but it is what it is.

@compyman added a great example of a use case for symlinks over here: https://github.com/pantsbuild/pants/pull/15211#issuecomment-1135155501

Some thoughts on how this might impact glob matching: `PosixFS` does already have a notion of [symlink awareness](https://github.com/pantsbuild/pants/blob/ebcf1a100fe9cfadac24df919ee14e16c1d0a923/src/rust/engine/fs/src/lib.rs#L354-L358), which currently only controls whether [`scandir` will ever return a symlink](https://github.com/pantsbuild/pants/blob/7e266c7770406110a25b5402b47cd8389b58cb68/src/rust/engine/fs/src/lib.rs#L467-L477). Glob...

@danxmoran : Would you mind re-running with `--streaming-workunits-level=trace`? Part of the issue appears to be that the relevant process isn't marked `level=LogLevel.DEBUG`: https://github.com/pantsbuild/pants/blob/4d9fe544663082c4c5816945bc3ac457c43a153c/src/python/pants/backend/shell/shell_command.py#L225-L234 ... I'll get a change out to...

Opened #16848 : it fixes the display aspect of this issue, but not the fact that capturing inputs/outputs took a long time (...which should become more evident with a `trace`-level...

From the additional `trace`-level trace that you provided offline, it looks like at least two things are happening: 1. [`eager_fetch_action_cache`](https://github.com/pantsbuild/pants/blob/221565246db4f31822ae63f6cc8cffb83dd6740a/src/rust/engine/process_execution/src/lib.rs#L784-L835) is taking a very long time to determine that the...