feat(changelog): support external commands for commit preprocessors (#86)

README.md

@@ -523,6 +523,19 @@ Examples:
 - `{ pattern = "([ \\n])(([a-f0-9]{7})[a-f0-9]*)", replace = "${1}commit # [${3}](https://github.com/orhun/git-cliff/commit/${2})"}`
   - Hyperlink bare commit hashes like "abcd1234" in commit logs, with short commit hash as description.
 
+Custom OS commands can also be used for modifying the commit messages:
+
+- `{ pattern = "foo", replace_command = "pandoc -t commonmark"}`
+
+This is useful when you want to filter/encode messages using external commands. In the example above, [pandoc](https://pandoc.org/) is used to convert each commit message that matches the given `pattern` to the [CommonMark](https://commonmark.org/) format.
+
+A more fun example would be reversing the each commit message:
+
+- `{ pattern = '.*', replace_command = 'rev | xargs echo "reversed: $@"' }`
+
+`$COMMIT_SHA` environment variable is set during execution of the command so you can do fancier things like reading the commit itself:
+
+- `{ pattern = '.*', replace_command = 'git show -s --format=%B $COMMIT_SHA' }`
 
 #### commit_parsers
 

git-cliff-core/src/command.rs

@@ -0,0 +1,81 @@
+use crate::error::Result;
+use std::io::{
+	Error as IoError,
+	ErrorKind as IoErrorKind,
+	Write,
+};
+use std::process::{
+	Command,
+	Stdio,
+};
+use std::str;
+use std::thread;
+
+/// Runs the given OS command and returns the output as string.
+///
+/// Use `input` parameter to specify a text to write to stdin.
+/// Environment variables are set accordingly to `envs`.
+pub fn run(
+	command: &str,
+	input: Option<String>,
+	envs: Vec<(&str, &str)>,
+) -> Result<String> {
+	let mut child = if cfg!(target_os = "windows") {
+		Command::new("cmd")
+			.args(&["/C", command])
+			.stdin(Stdio::piped())
+			.stdout(Stdio::piped())
+			.spawn()
+	} else {
+		Command::new("sh")
+			.envs(envs)
+			.args(&["-c", command])
+			.stdin(Stdio::piped())
+			.stdout(Stdio::piped())
+			.spawn()
+	}?;
+	if let Some(input) = input {
+		let mut stdin = child.stdin.take().ok_or_else(|| {
+			IoError::new(IoErrorKind::Other, "stdin is not captured")
+		})?;
+		thread::spawn(move || {
+			stdin
+				.write_all(input.as_bytes())
+				.expect("Failed to write to stdin");
+		});
+	}
+	let output = child.wait_with_output()?;
+	if output.status.success() {
+		Ok(str::from_utf8(&output.stdout)?.to_string())
+	} else {
+		Err(IoError::new(
+			IoErrorKind::Other,
+			format!("command exited with {:?}", output.status),
+		)
+		.into())
+	}
+}
+
+#[cfg(test)]
+mod test {
+	use super::*;
+	#[test]
+	#[cfg(target_family = "unix")]
+	fn run_os_command() -> Result<()> {
+		assert_eq!(
+			"eroc-ffilc-tig",
+			run("echo $APP_NAME | rev", None, vec![(
+				"APP_NAME",
+				env!("CARGO_PKG_NAME")
+			)])?
+			.trim()
+		);
+		assert_eq!(
+			"eroc-ffilc-tig",
+			run("rev", Some(env!("CARGO_PKG_NAME").to_string()), vec![])?.trim()
+		);
+		assert_eq!("testing", run("echo 'testing'", None, vec![])?.trim());
+		assert!(run("some_command", None, vec![]).is_err());
+		Ok(())
+	}
+}

git-cliff-core/src/commit.rs

@@ -1,3 +1,4 @@
+use crate::command;
 use crate::config::{
 	CommitParser,
 	CommitPreprocessor,
@@ -75,7 +76,7 @@ impl Commit<'_> {
 	pub fn process(&self, config: &GitConfig) -> Result<Self> {
 		let mut commit = self.clone();
 		if let Some(preprocessors) = &config.commit_preprocessors {
-			commit = commit.preprocess(preprocessors);
+			commit = commit.preprocess(preprocessors)?;
 		}
 		if config.conventional_commits.unwrap_or(true) {
 			if config.filter_unconventional.unwrap_or(true) {
@@ -109,17 +110,31 @@ impl Commit<'_> {
 
 	/// Preprocesses the commit using [`CommitPreprocessor`]s.
 	///
-	/// Modifies the commit [`message`] using regex.
+	/// Modifies the commit [`message`] using regex or custom OS command.
 	///
 	/// [`message`]: Commit::message
-	pub fn preprocess(mut self, preprocessors: &[CommitPreprocessor]) -> Self {
-		preprocessors.iter().for_each(|preprocessor| {
-			self.message = preprocessor
-				.pattern
-				.replace_all(&self.message, &preprocessor.replace)
-				.to_string();
-		});
-		self
+	pub fn preprocess(
+		mut self,
+		preprocessors: &[CommitPreprocessor],
+	) -> Result<Self> {
+		preprocessors.iter().try_for_each(|preprocessor| {
+			if let Some(text) = &preprocessor.replace {
+				self.message = preprocessor
+					.pattern
+					.replace_all(&self.message, text)
+					.to_string();
+			} else if let Some(command) = &preprocessor.replace_command {
+				if preprocessor.pattern.is_match(&self.message) {
+					self.message = command::run(
+						command,
+						Some(self.message.to_string()),
+						vec![("COMMIT_SHA", &self.id)],
+					)?;
+				}
+			}
+			Ok::<(), AppError>(())
+		})?;
+		Ok(self)
 	}
 
 	/// Parses the commit using [`CommitParser`]s.

git-cliff-core/src/config.rs

@@ -86,9 +86,11 @@ pub struct CommitParser {
 pub struct CommitPreprocessor {
 	/// Regex for matching a text to replace.
 	#[serde(with = "serde_regex")]
-	pub pattern: Regex,
+	pub pattern:         Regex,
 	/// Replacement text.
-	pub replace: String,
+	pub replace:         Option<String>,
+	/// Command that will be run for replacing the commit message.
+	pub replace_command: Option<String>,
 }
 
 /// Parser for extracting links in commits.

git-cliff-core/src/lib.rs

@@ -6,6 +6,8 @@ pub use glob;
 /// Export `regex` crate.
 pub use regex;
 
+/// Command runner.
+pub mod command;
 /// Git commit.
 pub mod commit;
 /// Config file parser.

git-cliff-core/tests/integration_test.rs

@@ -41,8 +41,9 @@ fn generate_changelog() -> Result<()> {
 		conventional_commits:  Some(true),
 		filter_unconventional: Some(true),
 		commit_preprocessors:  Some(vec![CommitPreprocessor {
-			pattern: Regex::new(r#"\(fixes (#[1-9]+)\)"#).unwrap(),
-			replace: String::from("[closes Issue${1}]"),
+			pattern:         Regex::new(r#"\(fixes (#[1-9]+)\)"#).unwrap(),
+			replace:         Some(String::from("[closes Issue${1}]")),
+			replace_command: None,
 		}]),
 		commit_parsers:        Some(vec![
 			CommitParser {

git-cliff/src/changelog.rs

@@ -192,8 +192,11 @@ mod test {
 				conventional_commits:  Some(true),
 				filter_unconventional: Some(false),
 				commit_preprocessors:  Some(vec![CommitPreprocessor {
-					pattern: Regex::new("<preprocess>").unwrap(),
-					replace: String::from("this commit is preprocessed"),
+					pattern:         Regex::new("<preprocess>").unwrap(),
+					replace:         Some(String::from(
+						"this commit is preprocessed",
+					)),
+					replace_command: None,
 				}]),
 				commit_parsers:        Some(vec![
 					CommitParser {