Overview

This is a rust attribute-like proc macro which reduces the amount of code required to call shell commands and parse the results.

It allows you to wrap a script in any language with strongly typed functions. The function's arguments are set as env variables and the result of the script is parsed either as a value or as an iterator.

Documentation

Repository

Examples

Basic

use shellfn::shell;
use std::error::Error;

#[shell]
fn list_modified(dir: &str) -> Result<impl Iterator<Item=String>, Box<Error>> { r#"
    cd $DIR
    git status | grep '^\s*modified:' | awk '{print $2}'
"# }

Different interpreter

use shellfn::shell;
use std::error::Error;

#[shell(cmd = "python -c")]
fn pretty_json(json: &str, indent: u8, sort_keys: bool) -> Result<String, Box<Error>> { r#"
import os, json

input = os.environ['JSON']
indent = int(os.environ['INDENT'])
sort_keys = os.environ['SORT_KEYS'] == 'true'
obj = json.loads(input)

print(json.dumps(obj, indent=indent, sort_keys=sort_keys))
 "# }

Usage

You can use the #[shell] attribute on functions that have:

• a body containing only one expression - a string literal representing the script to execute
• types that implement the .to_string() method
• return a value that is either void, T, Result<T, E>, impl Iterator<Item=T>, Result<impl Iterator<Item=T>> or Result<impl Iterator<Item=Result<T, E>>> with constrains:

T: FromStr,
<T as FromStr>::Err: StdError,
E: From<shellfn::Error<<T as FromStr>::Err>>,

• Details

The #[shell] attribute does the following:

1. Sets every argument as an env variable
2. Runs a shell command
3. Launches the command using std::process::Command
4. Depending on the return type, it may parse the output

Most of the steps can be adjusted:

• the default command is bash -c. You can change it using the cmd parameter:

#[shell(cmd = "python -c")]

• by default, the script is added as the last argument. You can change it using the special variable PROGRAM in the cmd parameter:

#[shell(cmd = "bash -c PROGRAM -i")]

• you can use env variables set from function's arguments in the cmd parameters in the same way as in the script:

#[shell(cmd = "python -m $MODULE")]
fn run(module: &str)

• if the return type is not wrapping some part of the result in Result, you may decide to suppress panics by adding the no_panic flag:

#[shell(no_panic)]

Following return types are currently recognized:

return type	flags	on parse fail	on error exit code	on spawn fail	notes
		-	panic	panic
	no_panic	-	nothing	nothing
()		-	panic	panic
()	no_panic	-	nothing	nothing
Result<(), E>		-	error	error
Result<(), E>	no_panic	-	error	error	1
T		panic	panic	panic	2
T	no_panic	panic	panic	panic	1,2
Result<T, E>		error	error	error	2
Result<T, E>	no_panic	error	error	error	1,2
Vec		panic	panic	panic
Vec	no_panic	skip	ignored	empty vec	3
Vec<Result<T, E>>		item error	panic	panic
Vec<Result<T, E>>	no_panic	item error	ignored	empty vec
Result<Vec, E>		panic	error	error
Result<Vec, E>	no_panic	skip	error	error
Result<Vec<Result<T, E1>>, E2>		item error	error	error
Result<Vec<Result<T, E1>>, E2>	no_panic	item error	error	error	1
impl Iterator<Item=T>		panic	panic	panic
impl Iterator<Item=T>	no_panic	skip	ignored	empty iter	3
impl Iterator<Item=Result<T, E>>		item error	panic	panic	3
impl Iterator<Item=Result<T, E>>	no_panic	item error	ignored	empty iter
Result<impl Iterator<Item=T>, E>		panic	ignored	error
Result<impl Iterator<Item=T>, E>	no_panic	skip	ignored	error
Result<impl Iterator<Item=Result<T, E1>>, E2>		item error	ignored	error
Result<impl Iterator<Item=Result<T, E1>>, E2>	no_panic	item error	ignored	error	1

Glossary:

action	meaning
panic	panics (.expect or panic!)
nothing	consumes and ignores error (let _ = ...)
error	returns error
skip	yields all successfuly parsed items, ignores parsing failures (filter_map)
empty iter/vec	returns empty iterator / vector
item error	when parsing fails, yields Err
ignored	ignores exit code, behaves in the same way for exit code 0 and != 0

Notes:

1. The no_panic attribute makes no difference
2. It reads all of stdout before producing any failures
3. It yields all items until it encounters an error or an exit code

Vector vs iterator

Variants with the Vec return type are very similar to the ones with impl Iterator. The key differences are:

• impl Iterator is only allocating one item at the time and yields it immediately after it is parsed, while Vec is reading output line by line but stores parsed output in the temporary Vec
• Vec is aware of exit code. When subprocess finishes with error, impl Iterator will stop yielding values while Vec will return error or panic

Contribution

All contributions and comments are more than welcome! Don't be afraid to open an issue or PR whenever you find a bug or have an idea to improve this crate.

License

MIT License

Copyright (c) 2017 Marcin Sas-Szymański

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.