[Toybox] About the sort -k-1 thing...

[Rob Landley]

I'm still subscribed to the coreutils mailing list because:

  https://lists.gnu.org/archive/html/coreutils/2023-08/msg00009.html

And recently there was:

  https://lists.gnu.org/archive/html/coreutils/2023-08/msg00035.html

So I did

  https://lists.gnu.org/archive/html/coreutils/2023-08/msg00048.html

And still need to do the "-k 1.-2" offsets part, but I'm out of brain to come up
with tests for them today, and squinting at the existing code there's no bounds
checking? I'm mean there is for the whole string, but it doesn't care that the
offset stays within the same key, which is fine until you try to figure out what
"from the end" logic looks like. (I guess it's ok to back up into previous keys
with a large enough negative value, which is now how the key indexes themselves
work?) Or if you can advance past the end of this key then why doesn't a
negative value back up from the START of this key into the previous key?

I'm not convinced I know what a negative offset should mean in this context. I
should ask on the coreutils list again since that's where this came from, but I
haven't got the spoons to venture back into GNU territory just now:

  https://www.youtube.com/watch?v=qXH5CD3O7Oc

Rob