Re: compadd --

[Bart Schaefer <schaefer@xxxxxxxxxxxxxxxx>]

On Oct 25,  5:29pm, Peter Stephenson wrote:
} Subject: Re: compadd --
}
} On Tue, 25 Oct 2011 09:04:45 -0700
} Bart Schaefer <schaefer@xxxxxxxxxxxxxxxx> wrote:
} > BUILTIN("compadd", 0, bin_compadd, 0, -1, 0, NULL, NULL)
} > 
} > says that compadd is going to do all its own option parsing.
} 
} Except it doesn't.  Are you saying we need this?

Oops, I'd forgotten about BINF_HANDLES_OPTS; I thought a NULL in the
opts slot meant there were no options and therefore the parser should
not remove any (including not removing "--").

Anyway that patch does appear to fix it.
 
} (Possibly for other completion functions?)

Well, having the generic option parser handle "--" was a change dating
from ... when?  Sometime after 4.2.x ... and if we missed this one, we
may have missed others.  But I'm having a hard time coming up with any
reasonable way to check other than going through it all by eye.