table tests need TLC

[pfuntner]

I was revisiting table/tests.py and noticed it was in bad need of work:

• It was using Python 2
• It wasn't using f-strings

I put in a bit of work and got rid of syntax errors (66d2dda) but it's claiming there are failures:

$ table/test.py 
Testing ls
2023-09-23 10:09:47,520 CRITICAL /ExtraVolume/repos/toys/test/table/test.py:29 xml -> xml mismatch
$

I took a look and I think it's working correctly:

$ diff <(table -i fixed -o xml < table/ls/original.txt | html) <(table -i fixed -o xml < table/ls/original.txt | table -i xml -o xml | html)
$ table -i fixed -o xml < table/ls/original.txt | html | head
<table>
  <row>
    <col00000000>perms
    </col00000000>
    <col00000001>l
    </col00000001>
    <col00000002>user
    </col00000002>
    <col00000003>group
    </col00000003>
$ 

However the expected output is different:

$ html < table/ls/xml/xml/noheadings.expected | head
<table>
  <row>
    <col00000005>datetime
    </col00000005>
    <col00000004>size
    </col00000004>
    <col00000006>name
    </col00000006>
    <col00000001>l
    </col00000001>

The columns are clearly in a different order... it shouldn't make a difference in the XML but it's throwing the test off. I actually prefer the new behavior so the expected output file needs to be updated. I know I periodically add features and tweaks to table.py - I even added something to not sort dictionary keys but it shouldn't have changed this (famous last words!).

[pfuntner]

I wonder if I need to regenerate all the test data. I could write a script to do that! I must have had a script to do it originally. I don't think I saved it, however.

[pfuntner]

This could be a dupe of #94