On 2017-11-15 at 22:44 +0100, martin f krafft wrote: > First, we need to debate to death whether it should be > > while sleep 1; do echo .; done > > or > > while echo .; do sleep 1; done > > ;) Hah. Clearly `while sleep N`, because that way killing the sleep propagates the death out of the loop, making it easier to clean up afterwards. My bikeshed should be blue. I realized afterwards that really, this is too frequent and also that I didn't think through what ssh's default buffering behavior might be without a PTY. The simplest fix is thus to just use `ssh -t` to get line-buffering and have each echo try to pass content to the client. Otherwise, >> dd if=/dev/zero bs=1500 count=1 << is likely to work, but at the cost of a chunk of unnecessary traffic. Really, sacrifice a PTY and let the systems all automatically decide to flush output on each line. -Phil _______________________________________________ openssh-unix-dev mailing list openssh-unix-dev@xxxxxxxxxxx https://lists.mindrot.org/mailman/listinfo/openssh-unix-dev