>>>>> On Fri, 20 Dec 2019 14:24:39 +0100, Robert Pluim <[email protected]> said:

>>>>> On Fri, 20 Dec 2019 13:40:12 +0100, Michael Albinus 
>>>>> <[email protected]> said:
    Michael> Apply the patch, and rerun the test. If you set tramp-verbose to 
10 in
    Michael> advance, there shall be a backtrace in the Tramp debug buffer, 
telling
    Michael> us at which point the quit signal has been throwed.

    Michael> (Or it runs w/o problems, and I'll be astonished :-)

    Robert> It still has the same problem, but the debug buffer has:

    Robert> 14:16:09.310254 tramp-accept-process-output (10) # 
    Robert>   backtrace()
    Robert>   tramp-signal-hook-function(quit nil)
    Robert>   accept-process-output(#<process *tramp/ssh 127.1*> nil nil t)
    Robert>   tramp-accept-process-output(#<process *tramp/ssh 127.1*>)
    Robert>   tramp-wait-for-regexp(#<process *tramp/ssh 127.1*> nil 
"\\(^\\|\0\\)[^#$\n]*///4b7473023f8bf973fb72feb160e879ca...")
    Robert>   tramp-wait-for-output(#<process *tramp/ssh 127.1*>)

And that plus the fact that there was a 'git log' process hanging
around on the remote machine indicated somebody was waiting for
something to happen that never would, which lead me to this in my
(remote) .gitconfig

    pager = less -FRXI

Once I unset that, everything started working fine. Magit works
because it passes '--no-pager' to git, which means this is a VC issue,
not a Tramp issue.

Thanks

Robert

Reply via email to