[2.x] test: Add test for sbt --client startup time (#8599)

* Add test for sbt --client startup time
* Skip startup time test on macOS due to slow hostname resolution
This commit is contained in:
calm 2026-01-21 17:00:00 -08:00 committed by GitHub
parent 2a5746cf6c
commit 37c6eb2cd9
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 43 additions and 0 deletions

View File

@ -138,4 +138,47 @@ object ExtendedRunnerTest extends BasicTestSuite:
}
()
}
// Test for issue #6485: Test `sbt --client` startup
// https://github.com/sbt/sbt/issues/6485
test("sbt --client startup time") {
if (isWindows || isMac) {
// Skip on Windows (sbtn behavior differs) and macOS CI (slow hostname resolution)
()
} else {
// First call starts the server if not running (warmup)
val warmup = sbtProcess("--client", "version").!
assert(warmup == 0, "Warmup sbt --client version failed")
// Measure startup time for sbt --client when server is already running
// Run multiple times and take the average to reduce variance
val iterations = 5
val times = (1 to iterations).map { _ =>
val start = System.nanoTime()
val exitCode = sbtProcess("--client", "version").!
val elapsed = (System.nanoTime() - start) / 1_000_000 // Convert to milliseconds
assert(exitCode == 0, "sbt --client version failed")
elapsed
}
val avgTime = times.sum / iterations
val maxTime = times.max
println(s"sbt --client startup times (ms): ${times.mkString(", ")}")
println(s"Average: ${avgTime}ms, Max: ${maxTime}ms")
// Cap at 2000ms to catch significant regressions while allowing for CI variance.
// The original issue #5980 mentioned ~200ms on developer machines in 2021,
// but CI runners are typically 2-3x slower than local development machines.
assert(
avgTime < 2000,
s"sbt --client startup time (${avgTime}ms average) exceeded 2000ms threshold"
)
// Cleanup: shutdown the server
val shutdown = sbtProcess("--client", "shutdown").!
assert(shutdown == 0, "Failed to shutdown sbt server")
}
()
}
end ExtendedRunnerTest