v1.1.0 (Nov 2, 2023)

In this release:

• #77: --safety-check flag for checking that the output of smlfmt is correct, i.e., that the result is valid SML that exactly matches the source except for whitespace.
  • Note: originally called --check but renamed to --safety-check in #92.
• #84, #86: allow passing input and output through stdin and stdout.
  • For example: echo "val x = 5 val y = 6" | ./smlfmt or cat test.sml | ./smlfmt > output.sml
• #87: --read-only flag, useful for checking for syntax errors
• #93: make install formula
• #92: --check flag, to check if a file has already been formatted. Useful for CI.
• Some bugfixes and other small improvements, e.g., 656885b 4d1952a

Also note that https://github.com/diku-dk/smlfmt.el offers Emacs integration for smlfmt.

Thanks @kopecs @p-ouellette @brandonspark @athas @ratsclub for your contributions!

v1.0.0 (Jan 10, 2023)

smlfmt is a custom parser and code formatter for Standard ML.

Features

• Full support for Standard ML according to the formal definition.
• Fast -- see performance results below.
• Usable for large projects.
  • Whole-project formatting via .mlb files using MLton conventions.
• Helpful error messages with visual code references and syntax highlighting.
• Support for various SuccessorML syntax extensions.
• Configurable indentation size and maximum width.

Performance

smlfmt is fast enough for live reformatting on large individual files, and can reformat large projects in seconds.

Here are measurements for smlfmt --preview-only INPUT > out on my MacBook Air (M2, 2022).

Input	Size (LoC)	Time
src/prettier-print/PrettierExpAndDec.sml	1024	31 ms
src/base/PrettyTabbedDoc.sml	1349	29 ms
src/parse/ParseExpAndDec.sml	1363	28 ms
MLton elaborate-core.fun	3942	104 ms
MLton x86-allocate-registers.fun	11034	216 ms
Entire MLton source code (mlton.mlb)	160000 (approx.)	6.6s

(Timings are averages reported by hyperfine.)