Re: [RFC-PATCHv6 4/4] pathspec: allow querying for attributes

Stefan Beller [off-list ref] writes:

I'll just drop support for values
in the first series.

I do not think an exact string match to support :(attr:eol=crlf) is
so bad.  The "crazy stuff" aka over-engineering is when it goes
beyond that, e.g. 'eol is set to one of these values"