Re: [PATCH 3/6] completion: return the index of found word from __git_find_on_cmdline()

From: SZEDER Gábor <szeder.dev@gmail.com>
Date: Oct 18, 2019, 14:37 UTC
Message-ID: <20191018143728.GC29845@szeder.dev>
In-Reply-To: <CAPig+cTLCTPtUWHKVBQEaP9GkrmrnHPwfef8KnktOSqYQY-jPA@mail.gmail.com>

On Thu, Oct 17, 2019 at 01:52:27PM -0400, Eric Sunshine wrote:

Show 18 quoted lines

> >  __git_find_on_cmdline ()
> >  {
> > -       local word c=1
> > +       local word c=1 show_idx
> > +
> > +       while test $# -gt 1; do
> > +               case "$1" in
> > +               --show-idx)     show_idx=y ;;
> > +               *)              return 1 ;;
> 
> Should this emit an error message to aid a person debugging a test
> which fails on a call to __git_find_on_cmdline()? For instance:
> 
>     echo "unrecognized option/argument: $1" >&2
>     return 1
>     ;;
> 
> or something...

When debugging the completion script I frequently resort to 'echo >&2 "<msg>"', for lack of better options. However, I intentionally did not add an error message like that here, or in any similar option parsing loops before, because due to a bug it might spew such a message to standard error during regular completion (i.e. not during debugging).

And printing anything to standard error during completion is inherently bad: it disrupts the command line, can't be deleted (you hit backspace, and in the terminal it looks as if the error message was deleted, but in reality it's the command you've already entered that gets deleted), and the user is eventually fored to Ctrl-C and start over most of the time. Well, at least I always end up hitting Ctrl-C and start over. Remaining silent about the unrecognized option is in my opinion better, because then the completion script usually does nothing, and Bash falls back to filename completion. Yeah, that's not ideal, but at least the user can easily correct it and finish entering the command.