Re: [PATCH] gitk: fix --all behavior combined with --not

> Junio C Hamano <gitster@pobox.com> writes:
> 
> > Heiko Voigt <hvoigt@hvoigt.net> writes:
> >
> >> In commit 4d5e1b1319 ("gitk: Show detached HEAD if --all is specified",
> >> 2014-09-09) the intention was to have detached HEAD shown when the --all
> >> argument is given.
> >
> > The "do we have --all?" test added by that old commit is not quite
> > satisfying in the first place.  E.g. we do not check if there is a
> > double-dash before it.  This change also relies on an ancient design
> > mistake of allowing non-dashed options before a dashed one, adding
> > more to dissatisfaction by making a future change to correct the
> > design mistake harder.
> 
> Actually, I do not think this patch is a good idea.
> 

> 
> As the code is _already_ finding the _exact_ location on the command
> line where "--all" appears, I think you can go one step further and
> make sure you insert the "HEAD" immediately after "--all", as that
> exactly matches what you (and the ancient 4d5e1b1319) are trying to
> achieve: pretend as if "--all" always include "HEAD", even when it
> is detached.
> 
> This is orthogonal to the question I posed in my earlier reply
> (i.e. "we found --all; is it really a 'give me all refs' request
> given by the user, or something else (is it an argument to another
> option, like "--grep '--all'", or is it pathspec after '--'), but
> assuming that we have reliably found the "--all" on the command line
> the user meant as "give me all refs", I think inserting HEAD
> immediately after that location would be the right solution.  It is
> incorrect to unconditionally append as your original example shows,
> but it is equally incorrect to unconditionally prepend.