Re: PATCH: Re: 3.1.6-dev-22

[Adam Spiers <adam@xxxxxxxxxx>]

Sven Wischnowsky (wischnow@xxxxxxxxxxxxxxxxxxxxxxx) wrote:
> Andrej Borsenkow wrote:
> > The decsription of compstate[insert] :
> > 
> > ...
> >           On exit it may be set to any of the values above (where
> >           setting it to the empty string is the same as unsetting it),
> >           or to a number, in which case the match whose number is given
> >           will be inserted into the command line.  It may also be set
> >           to a string of the form `GROUP:MATCH' which specifies a match
> >           from a group of matches to be inserted, counting from 1
> >           upwards (e.g. `2:4' specifies the fourth match of the second
> >           group).  Negative numbers count backward from the last match
> > ... etc and the next paragraph
> > 
> > Now, either I'm completely blind, or this is the _only_ place where
> > "number of group" is mentioned. This was there for a long time; I
> > presume, originally it was for sorted/unsorted matches - but
> > 
> > - either it should be described, where these group numbers come from
> > - or, better yet, the name of group should be used.
> 
> We have six name spaces for group names.
> 
> The whole thing comes from a time when we were thinking about ways to
> get information about the matches already added. Then I was thinking
> about using it for the stuff that is now done with _next_tags. I
> didn't expect so much consistency and control in the shell code then.
> 
> So, I would like to make the question: should we remove it? I would
> only comment it out in the C-code just in case we get a way to access
> matches added some day (not for a long time, I think).
> 
> Actually, I was already tempted to remove it more than once and I
> don't think it would cause much harm (or any at all) -- noone has
> written another example completion system and the one we have doesn't
> use this feature.

Doesn't history-complete-word need this feature?