On Tue, Jan 15, 2019 at 11:40:28AM -0800, Jonathan Tan wrote: > A subsequent patch will teach struct packet_reader a new field that, if > set, instructs it to interpret read data as multiplexed. This will > create a dependency from pkt-line to sideband. > > To avoid a circular dependency, split recv_sideband() into 2 parts: the > reading loop (left in recv_sideband()) and the processing of the > contents (in demultiplex_sideband()), and move the former into pkt-line. > This reverses the direction of dependency: sideband no longer depends on > pkt-line, and pkt-line now depends on sideband. In the last couple of days I noticed occasional but frequent failures in the test 'leading space' in 't5409-colorize-remote-messas.sh' on 'pu' and on this topic. Bisect suggests that this patch is the culprit, but of course bisecting an occasional test failure can't be completely trusted. The trace output of the failing test looks like this: + git --git-dir child/.git -c color.remote=always push -f origin HEAD:refs/heads/leading-space + cat output remote: error: error remote: ERROR: also highlighted remote: hint: hint remote: hinting: not highlighted remote: success: success remote: warning: warning remote: prefixerror: error remote: remote: error: leading space remote: remote: Err To /<...>/trash directory.t5409-colorize-remote-messages.stress-4/. * [new branch] HEAD -> leading-space + test_decode_color + awk < ... snip enormous awk script ... > + grep <BOLD;RED>error<RESET>: leading space decoded error: last command exited with $?=1 not ok 6 - leading space Notice how that "error: leading space" lines up with the other messages. On 'master' that line looks like this: remote: error: leading space
