git-fmt

(Side note: the formatting component of this project will eventually be split out and named omnifmt.)

git-fmt was created to make prettifying code easy. It adds a custom (easy to use) command to Git that formats code through external pretty-printers.

Formatted code is:

• Easier to write: never worry about minor formatting concerns while hacking away.
• Easier to read: when all code looks the same you need not mentally convert others' formatting style into something you can understand.
• Easier to maintain: mechanical changes to the source don't cause unrelated changes to the file's formatting; diffs show only the real changes.
• Uncontroversial: never have a debate about spacing or brace position ever again.

(Bullet points taken from https://blog.golang.org/go-fmt-your-code.)

Installing

Installing git-fmt is easiest done using either stack (recommended) or Cabal.

Using stack:

stack install git-fmt
export PATH=$PATH:~/.local/bin

Using Cabal:

cabal-install git-fmt
export PATH=$PATH:~/.cabal/bin

Usage

The git-fmt binary wraps external pretty-printers and provides an interface for selecting files to prettify and performing dry-runs.

The basics:

git-fmt operates only on tracked git files (thus it implicitly respects the .gitignore). By default it operates on files in the index (i.e., --operate-on head). It is possible to operate on all tracked files (--operate-on-tracked) or on a specific reference (--operate-on REF). The REF argument is passed directly into git diff REF --name-only, so you can even play with ranges such as master....

Passing arguments to git-fmt will narrow down the operation files. For example, git-fmt --operate-on-tracked src/ will format all tracked files under src/ and git-fmt --operate-on head src/ will format all files in the index under src/.

git-fmt can run in two different modes, normal and dry-run. Normal mode (--mode normal) writes to (prettifies) all ugly files. Dry-run mode (--mode dry-run) outputs the ugly file paths to stdout.

NB: a third mode, patch, is in the making, see #23 for details.

For the more uncommon usage options, have a read of git fmt -h.

Configuration

Configuration is done via an .omnifmt.yaml file in the git repository. The file contains a list of programs that link extensions to a prettifying command, e.g.,

haskell:
    extensions: ["hs", "lhs"]
    command:    "stylish-haskell {{input}} > {{output}}"

javascript:
    extensions: ["js"]
    command:    "js-beautify -f {{input}}"

json:
    extensions: ["json"]
    command:    "json_pp"

ruby:
    extensions: ["rb"]
    command:    "ruby-beautify"

Each command declares how to read the input file and how to write to the output file. If the input variable is omitted, the file contents are fed to the command through stdin. Likewise if the output variable is omitted, the pretty contents are read from stdout. The output file is used to compare whether the original was pretty or ugly before writing to it.

The extensions field is pretty self explanatory, but if you use the same extension more than once then precedence goes to the program defined first.

Examples

See the docs/example-configs/ directory for some common pretty-printers and their corresponding omnifmt config (pull requests are welcome for adding more). Just don't forget to actually call the config file .omnifmt.yaml!

NB: I haven't tested them fully, be careful in case one is buggy.