Conflicts with CLICOLOR_FORCE

[defjaf]

Conflicts with CLICOLOR_FORCE

[scop]

This is the first time I've heard of CLICOLOR_FORCE and I don't have a system to test it with, but from what I understand I think enabling it globally is asking for trouble, and the proper way to use it is selectively where needed. I'm inclined to close this with the workaround being "just don't do it".

Current bash-completion still has the have() function but it's not really necessary any more nor used by bash-completion, it exists only for backwards compatibility. It's likely that the above issue still exists with current versions though. In any case, I suggest upgrading bash-completion to an up to date version before reporting issues here, it's extremely unlikely that any work on earlier versions than the current 2.x line will be done.

[akinomyoga]

I think this can also be closed. I agree that we shouldn't set the variable globally. That would break many shell scripts, and the issue is not specific to bash-completion.

I think a realistic usage for such a variable is to use it with an alias (alias ls='CLICOLOR_FORCE=1 ls'), a shell function (myls() { local -x CLICOLOR_FORCE=1; ls "$@"; }), or an executable script (#/usr/bin/env bash\nexport CLIPCOLOR_FORCE=1\nls "$@").