# git-completion.bash should not add a space after a ref

2 messages from 2013-01-12 to 2013-01-13. Participants: Manlio Perillo, Junio C Hamano.
Thread: https://gitlist.dev/t/32608

## Manlio Perillo, 2013-01-12 18:35

Subject: git-completion.bash should not add a space after a ref
Message-ID: <50F1AD0F.7080503@gmail.com>
URL: https://gitlist.dev/e/50F1AD0F.7080503%40gmail.com

```
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi.

This is not really a bug, but a small usability problem.

When completing a reference, Bash will add a space after the reference name.

As an example in:

    $git show master<TAB>

The problem is that an user may want to show a tree or blog object from
master:

    $git show master:git.c


A possible solution is to define a new __gitcomp_nospace function and
use it where appropriate.

Probably the __gitcomp_nospace should be used when git_complete_file is
called, and __gitcomp_nl should be used when __git_complete_revlist  is
called, but I'm not sure.

P.S.:
it seems that __gitcomp_nl is **never** called with 4 arguments.



Regards  Manlio
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAlDxrQ8ACgkQscQJ24LbaURHmACfRXoM+uEVDgFUtZFzUcPC5oSZ
FGsAnAxQf+SN7GrNljxU1io4IuayHmed
=JRVU
-----END PGP SIGNATURE-----

```

## Junio C Hamano, 2013-01-13 23:16

Subject: Re: git-completion.bash should not add a space after a ref
Message-ID: <7vk3rghdwv.fsf@alter.siamese.dyndns.org>
URL: https://gitlist.dev/e/7vk3rghdwv.fsf%40alter.siamese.dyndns.org
In-Reply-To: <50F1AD0F.7080503@gmail.com>

```
Manlio Perillo <manlio.perillo@gmail.com> writes:

> This is not really a bug, but a small usability problem.
>
> When completing a reference, Bash will add a space after the reference name.
>
> As an example in:
>
>     $git show master<TAB>
>
> The problem is that an user may want to show a tree or blog object from
> master:
>
>     $git show master:git.c

Or the user may want to see only changes to the documentation, e.g.

	$ git show master Documentation/

so removing a SP is a regression in the other way around.

Given that "refer to an object in a tree" is a much less often used
operation, I have a feeling that the current behaviour happens to
make a good trade-off between these two conflicting purposes that
cannot be satisfied both at the same time.

```