Tests failing on clean checkout from master.

[mindhog]

I'm seeing test failures on a clean checkout from master (commit cff343c). To be fair, I can't entirely rule out my vagrant setup. I had to disable UDP and switch to the the "peru/ubuntu-20.04-server-amd64" box, as the hashicorp box defined for ubuntu in tmux test doesn't work with the libvirt provider.

But the diff I'm seeing suggests a change in the file format:

< pane  0       0       :bash   1       :*      0       :/tmp   1       bash    :
< pane  blue    0       :vim    0       :       0       :/tmp   1       vim     :vim foo.txt
< pane  blue    1       :man    0       :-      0       :/tmp   0       bash    :
< pane  blue    1       :man    0       :-      1       :/usr/share/man 1       man     :man echo
< pane  blue    2       :bash   1       :*      0       :/tmp   1       bash    :
< pane  red     0       :bash   0       :       0       :/tmp   1       bash    :
< pane  red     1       :bash   0       :-Z     0       :/tmp   0       bash    :
< pane  red     1       :bash   0       :-Z     1       :/tmp   0       bash    :
< pane  red     1       :bash   0       :-Z     2       :/tmp   1       bash    :
< pane  red     2       :bash   1       :*      0       :/tmp   0       bash    :
< pane  red     2       :bash   1       :*      1       :/tmp   1       bash    :
< pane  yellow  0       :bash   1       :*      0       :/tmp/bar       1       bash    :
< window        0       0       1       :*      ce9e,200x49,0,0,1
< window        blue    0       0       :       ce9f,200x49,0,0,2
< window        blue    1       0       :-      178b,200x49,0,0{100x49,0,0,3,99x49,101,0,4}
< window        blue    2       1       :*      cea2,200x49,0,0,5
< window        red     0       0       :       cea3,200x49,0,0,6
< window        red     1       0       :-Z     135b,200x49,0,0[200x24,0,0,7,200x24,0,25{100x24,0,25,8,99x24,101,25,9}]
< window        red     2       1       :*      db81,200x49,0,0[200x24,0,0,10,200x24,0,25,11]
< window        yellow  0       1       :*      6781,200x49,0,0,12
---
> pane  0       0       1       :*      0       localhost       :/vagrant       0       bash    :
> pane  0       0       1       :*      1       localhost       :/tmp   1       bash    :
> pane  blue    0       0       :       0       localhost       :/tmp   1       bash    :
> pane  blue    1       0       :-      0       localhost       :/tmp   0       bash    :
> pane  blue    1       0       :-      1       localhost       :/usr/share/man 1       bash    :
> pane  blue    2       1       :*      0       localhost       :/tmp   1       bash    :
> pane  red     0       0       :       0       localhost       :/tmp   1       bash    :
> pane  red     1       0       :-      0       localhost       :/tmp   0       bash    :
> pane  red     1       0       :-      1       localhost       :/tmp   0       bash    :
> pane  red     1       0       :-      2       localhost       :/tmp   1       bash    :
> pane  red     2       1       :*      0       localhost       :/tmp   0       bash    :
> pane  red     2       1       :*      1       localhost       :/tmp   1       bash    :
> pane  yellow  0       1       :*      0       localhost       :/tmp/bar       1       bash    :
> window        0       0       :bash   1       :*      fbd2,200x49,0,0[200x2,0,0,0,200x46,0,3,1]       on
> window        blue    0       :bash   0       :       ce9f,200x49,0,0,2       on
> window        blue    1       :bash   0       :-      d0e3,200x49,0,0[200x15,0,0,3,200x33,0,16,4]     on
> window        blue    2       :bash   1       :*      cea2,200x49,0,0,5       on
> window        red     0       :bash   0       :       cea3,200x49,0,0,6       on
> window        red     1       :bash   0       :-      c611,200x49,0,0[200x11,0,0,7,200x11,0,12,8,200x25,0,24,9]       on
> window        red     2       :bash   1       :*      e441,200x49,0,0[200x15,0,0,10,200x33,0,16,11]   on
> window        yellow  0       :bash   1       :*      6781,200x49,0,0,12      on

I'm wondering if restore_file.txt needs an update?

[mindhog]

I've confirmed that the tests do a lot better with the last revision restore_file.txt was changed at (fee390d) -- "restore" passes, there is a one-line diff in the save test due to a difference in the current directory for the "man" command, which seems like it might reasonably be due to a difference in the box configuration.