Add support for targetless formatters #16480
Assignees
Labels
Comments
thejcannon
added a commit
that referenced
this issue
Aug 16, 2022
…ng) (#16535) This is prepwork for dissolving update-build-files from #13504, as well as #16480. No semantic changes have been made, just renaming and refactoring. Note that FmtResult hasn't been renamed, as it isn't target-specific. Commits are useful to review smile [ci skip-rust] [ci skip-build-wheels]
cczona
pushed a commit
to cczona/pants
that referenced
this issue
Sep 1, 2022
…ng) (pantsbuild#16535) This is prepwork for dissolving update-build-files from pantsbuild#13504, as well as pantsbuild#16480. No semantic changes have been made, just renaming and refactoring. Note that FmtResult hasn't been renamed, as it isn't target-specific. Commits are useful to review smile [ci skip-rust] [ci skip-build-wheels]
|
Thanks @thejcannon!!!! Epic |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Similar to #14102, we want to support target-less formatters. This dramatically lowers the barrier to entry for both plugin authors and users, especially for things like SQL formatters that are unlikely to have other Pants metadata.
The main challenge is our parallelization scheme: within a language, formatters must run sequentially. But across language groups, the groups should run in parallel. In other words, Python's Black can run at the same time as Google Java Format, but Black and Isort cannot.
We need a mechanism for targetless and target-aware formatters to express that they both belong to the same language. The common value for both is the underlying source file path.
The text was updated successfully, but these errors were encountered: