the current implementation is sub optimal as it creates a new queue element for each byte read from the process.
the new implementation queues arrays of bytes instead, resulting in fewer queue elements and in avoiding byte -> Byte -> byte conversions. this also reduces GC pressure in the Eclipse IDE process.
I also implemented read(final byte[] buf, final int off, final int len) which improves bulk read performance.
the current implementation is sub optimal as it creates a new queue element for each byte read from the process. the new implementation queues arrays of bytes instead, resulting in fewer queue elements and in avoiding
byte -> Byte -> byte
conversions. this also reduces GC pressure in the Eclipse IDE process.I also implemented
read(final byte[] buf, final int off, final int len)
which improves bulk read performance.