On 11 January 2018 at 23:34, Michael Harding <[email protected]> wrote:
> No, the command and cms stages are documented to write all command output > before writing the return code to the secondary stream, which is easily > verified. > The stall occurs because command was stuck writing the return code before > it posted eof on its primary. A simple elastic stage on the second stream > takes care of that. > My apologies, you're right on the return code following the output. We don't get eof on the primary yet until we have consumed the input record. Which only happens after writing the return code. I would suggest "faninany" rather than "elastic" unless you really want all return codes following all command output... Rob
