In order to keep the design of
treefmt simple, we support only formatters which adhere to a certain standard. This document outlines this standard. If the formatter you would like to use doesn't comply with the rules, it's often possible to create a wrapper script that transforms the usage to match the specification.
In this design, we rely on
treefmt to do the tree traversal, and only invoke
the code formatter on the selected files.
In order for the formatter to comply to this spec, it MUST follow the following rules:
1. Files passed as arguments
In order to be integrated to
treefmt's workflow, the formatter's CLI must adhere to the following specification:
<command>is the name of the formatting tool.
[options]is any number of flags and options that the formatter accepts.
[...<files>]is one or more files given to the formatter for processing.
It SHOULD processes only the specified files. Files that are not passed SHOULD never be formatted.
2. Write to changed files
Whenever there is a change to the code formatting, the code formatter MUST write to the changes back to the original location.
If there is no changes to the original file, the formatter MUST NOT write to the original location.
The code formatter SHOULD be indempotent. Meaning that it produces stable outputs.
We expect the formatter to be reliable and not break the semantic of the formatted files.