Functional Kotlin friendly way to create external system processes by leveraging:
- Kotlin coroutines
- the powerful but convoluted ProcessBuilder
repositories {
mavenCentral()
}
dependencies {
// Check the 🔝 maven central badge 🔝 for the latest $kotlinProcessVersion
implementation("com.github.pgreze:kotlin-process:$kotlinProcessVersion")
}
Or in your kotlin script:
@file:DependsOn("org.jetbrains.kotlinx:kotlinx-coroutines-core:1.8.1")
@file:DependsOn("com.github.pgreze:kotlin-process:$kotlinProcessVersion")
Starts a program and prints its stdout/stderr outputs to the terminal:
import com.github.pgreze.process.process
import kotlinx.coroutines.runBlocking
runBlocking {
val res = process("echo", "hello world")
check(res.resultCode == 0)
// By default process output is displayed in the console.
check(res.output == emptyList<String>())
}
The next step would probably be to capture the output stream, in order to process some data from our own-made script:
val output = process(
"./my-script.sh", arg1, arg2,
// Capture stdout lines to do some operations after
stdout = Redirect.CAPTURE,
// Default value: prints to System.err
stderr = Redirect.PRINT,
).unwrap() // Fails if the resultCode != 0
// TODO: process the output
println("Success:\n${output.joinToString("\n")}")
Notice that if you want to capture both stdout and stderr, there will be no way to differentiate them in the returned output:
val res = process(
"./long-and-dangerous.sh", arg1, arg2,
// Both streams will be captured,
// preserving their orders but mixing them in the given output.
stdout = Redirect.CAPTURE,
stderr = Redirect.CAPTURE,
// Allows to consume line by line without delay the provided output.
consumer = { line -> TODO("process $line") },
)
println("Script finished with result=${res.resultCode}")
println("stdout+stderr:\n" + res.output.joinToString("\n"))
It's also possible to redirect an output stream to a file, or manually by consuming a Flow instance.
import com.github.pgreze.process.Redirect
import com.github.pgreze.process.process
import java.io.File
import java.util.Collections
import kotlinx.coroutines.flow.toList
import kotlinx.coroutines.runBlocking
val errLines = Collections.synchronizedList(mutableListOf<String>())
val res = process(
"./my-script.sh", arg1, arg2,
// You can save the execution result in a file,
stdout = Redirect.ToFile(File("my-input.txt")),
// If you want to handle this stream yourself,
// a Flow<String> instance can be used.
stderr = Redirect.Consume { flow -> flow.toList(errLines) },
)
The last but not least, you can just silence a stream with Redirect.SILENT 😶
Several other options are available to control the script environment:
import com.github.pgreze.process.InputSource
import java.io.File
val res = process(
"./my-script.sh",
// Provides the input as a string, similar to:
// $ echo "hello world" | my-script.sh
stdin = InputSource.fromString("hello world"),
// Inject custom environment variables:
env = mapOf("MY_ENV_VARIABLE" to "42"),
// Override the working directory:
directory = File("./a/directory"),
)
There are other ways to provide the process input:
// From a file:
process(
"./my-script.sh",
stdin = InputSource.FromFile(File("my-input.txt")),
)
// From an InputStream:
process(
"./my-script.sh",
stdin = InputSource.fromInputStream(myInputStream)),
)
// Manually by using the raw OutputStream:
process(
"./my-script.sh",
stdin = InputSource.FromStream { out: OutputStream ->
out.write("hello world\n".toByteArray())
out.flush()
},
)