Re: [PATCH 5/5] pathspec: record labels

From: Junio C Hamano <gitster@pobox.com>
Date: May 16, 2016, 18:52 UTC
Message-ID: <xmqq8tz93kf9.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <CAGZ79kbN_pzAVeAa-St_KRs60SbURTfKKP0v+do_+MK7orTkTg@mail.gmail.com>

Stefan Beller <sbeller@google.com> writes:

> So "warn and ignore" for data from .gitattributes and die for
> commandline arguments? That makes sense.

Yes.

On the "command line" front, because we may want to give different meanings to these two entries in the future:

        :(label=-doc)Documentation/
	:(label=!doc)Documentation/

we should diagnose -doc (FALSE) as an error, not treating it as the same as !doc (UNSET). And we should warn and ignore -doc (FALSE) in .gitattributes. Yes, ignoring it would be more or less equivalent to treating it as UNSET, but because we may use -doc (FALSE) for a better purpose later, we should still warn.

Show 9 quoted lines

> Ok, so here is the warn-and-ignore code:
>
>
>         if (ATTR_TRUE(check.value))
>                 ret = 1; /* has all the labels */
>         else if (ATTR_FALSE(check.value)) {
>                 warning(_("Path '%s': Label must not be false. Treat
> as if no label was set"), path);
>                 ret = 0;

s/Treat as if .../The -label may be used differently in future versions of Git, so do not use it/;

But if we are going in the direction of :(attr:crlf=auto), all this discussion is moot, isn't it? I haven't formed a firm opinion on this, but it sure does sound tempting, doesn't it?