Clarify behavior of empty CLICOLOR and CLICOLOR_FORCE

[bradking]

The specification does not currently say anything about semantics for CLICOLOR= and CLICOLOR_FORCE=. The example code for CLICOLOR_FORCE implies that CLICOLOR_FORCE= should be equivalent to CLICOLOR_FORCE=1, but IMO that is confusing. An empty setting should be equivalent to no setting at all.

[jhasse]

I would prefer empty meaning "unset", too. Unfortunately that's not how Apple implemented it :(

Apple is probably the biggest supporter of this "spec". Maybe we can ask them somehow? Or just clarify the documentation.

[bradking]

I suggest we update the documentation to state that behavior for the empty case is unspecified so that tools can do either and users are encouraged to always set it with 0 or 1 explicitly.

[mawillers]

[...] behavior for the empty case is unspecified [...]

Would be fine for me.

$ CLICOLOR_FORCE=0 ls | cat
[colored output]

Apple is really ignoring the value of that variable altogether? That goes quite against the whole idea...

[lionkor]

Hate to be that guy, especially after this long, but CLICOLOR_FORCE was only ever meant to force ENABLE colors, not force DISABLE colors. It exists because while CLICOLOR enables colors, it should still not output any colors if the output is not a TTY. Therefore, _FORCE was invented.