The JVM implementation for the stdout and stderr streams for Processes generated by ProcessBuilder.spawn currently use a chunk size of 8192. While this is probably fine for most use cases I've noticed that it can limit throughput drastically when large amounts of data are transferred this way.
It would be helpful, if the chunk size would be a parameter, either of ProcessBuilder or by having a variant of stderr that takes a chunkSize parameter like
def stdout(chunkSize: Int = 8192): Stream[F, Byte]
The JVM implementation for the stdout and stderr streams for Processes generated by ProcessBuilder.spawn currently use a chunk size of 8192. While this is probably fine for most use cases I've noticed that it can limit throughput drastically when large amounts of data are transferred this way.
It would be helpful, if the chunk size would be a parameter, either of ProcessBuilder or by having a variant of stderr that takes a chunkSize parameter like
def stdout(chunkSize: Int = 8192): Stream[F, Byte]