Mosh regression between 10.x and 11-stable
cgull at glup.org
Thu Aug 11 23:51:46 UTC 2016
That's interesting (not in a good way).
Next thing to try:
mosh --server='/usr/local/bin/mosh-server 2>/dev/null' peter at VPS
(with an unmodified mosh-server)
Mosh prints the 'MOSH CONNECT ...' message on stdout, then forks. The
parent exits immediately, and the child prints verbose and copyright
info to stderr. I'm wondering if the race is that the parent and child
writes appear mixed together, corrupting the expected message.
You might try adding a 'print;' in the while loop that digests this
input in the mosh script. That'll tell us whether the script is not
getting 'MOSH CONNECT...' at all, or if it's corrupted. You'll have to
run mosh inside /usr/bin/script to capture that.
On 08/11/16 03:53 PM, Peter Jeremy wrote:
> On 2016-Aug-11 12:30:23 -0400, John Hood <cgull at glup.org> wrote:
>> I still can't reproduce this on 3 different 11.0-BETA4 servers and a
>> variety of clients and networks. Can you try and identify a more
>> portable repro or at least figure out why it fails on your system?
>> Please try applying this patch, too. It's a shot in the dark, though.
> That patch seems to fix the problem I'm seeing. Not waiting for output
> to drain is consistent with the symptoms I'm seeing, though I have no
> idea why only my Linux client is affected.
More information about the freebsd-current