git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [RFC PATCH] bash completion: complete refs for git-grep
@ 2009-10-06 10:08 Thomas Rast
  2009-10-06 15:45 ` Shawn O. Pearce
  0 siblings, 1 reply; 6+ messages in thread
From: Thomas Rast @ 2009-10-06 10:08 UTC (permalink / raw)
  To: git; +Cc: Shawn O. Pearce

Attempt ref completion once we have seen a regular expression, to help
the user with entering the <treeish> arguments.

Signed-off-by: Thomas Rast <trast@student.ethz.ch>
---

The use-case for this is actually a bit protracted but came up on IRC
yesterday: pasky asked if there was a simple way to grep through a
certain file in all refs.

Turns out git-grep already has half the required support: when given a
series of refs, it prefixes the matches with the ref, so the output is
already in a useful format.

Sadly it does not appear to support --all, --branches or similar
(which would be material for a separate patch).  But bash completion
can step in here: with M-*, it can expand all possible completions for
the current word onto the command line.

This is still RFC because, as you can see in the code below, I tried
to avoid completing at all while the user still needs to supply a
regex.  Sadly, bash turns the COMPREPLY=() into filename completion
anyway.  Is there a way to prevent this?  Otherwise the regex
complication should probably just go away and we can complete refs
always.


 contrib/completion/git-completion.bash |   18 ++++++++++++++++++
 1 files changed, 18 insertions(+), 0 deletions(-)

diff --git a/contrib/completion/git-completion.bash b/contrib/completion/git-completion.bash
index 6fd7e1d..c8cced6 100755
--- a/contrib/completion/git-completion.bash
+++ b/contrib/completion/git-completion.bash
@@ -1048,6 +1048,24 @@ _git_grep ()
 		return
 		;;
 	esac
+
+	local i c=1 have_regex=""
+	while [ $c -lt $COMP_CWORD ]; do
+		i="${COMP_WORDS[c]}"
+		case "$i" in
+		-e) ;;
+		-e*) have_regex="$c" ; break ;;
+		-*) ;;
+		*) have_regex="$c"; break ;;
+		esac
+		c=$((++c))
+	done
+
+	if [ -n "$have_regex" ]; then
+		__gitcomp "$(__git_refs)"
+		return
+	fi
+
 	COMPREPLY=()
 }
 
-- 
1.6.5.rc2.251.g34f85

^ permalink raw reply related	[flat|nested] 6+ messages in thread

* Re: [RFC PATCH] bash completion: complete refs for git-grep
  2009-10-06 10:08 [RFC PATCH] bash completion: complete refs for git-grep Thomas Rast
@ 2009-10-06 15:45 ` Shawn O. Pearce
  2009-10-07 15:27   ` Thomas Rast
  0 siblings, 1 reply; 6+ messages in thread
From: Shawn O. Pearce @ 2009-10-06 15:45 UTC (permalink / raw)
  To: Thomas Rast; +Cc: git

Thomas Rast <trast@student.ethz.ch> wrote:
> This is still RFC because, as you can see in the code below, I tried
> to avoid completing at all while the user still needs to supply a
> regex.  Sadly, bash turns the COMPREPLY=() into filename completion
> anyway.  Is there a way to prevent this?

Not that I know of.  You can turn off default filename completion
when you register the completion function, but that then breaks
like every other git command for completion support because a lot
of them do want to complete filenames.

> +	local i c=1 have_regex=""
> +	while [ $c -lt $COMP_CWORD ]; do
> +		i="${COMP_WORDS[c]}"
> +		case "$i" in
> +		-e) ;;
> +		-e*) have_regex="$c" ; break ;;
> +		-*) ;;
> +		*) have_regex="$c"; break ;;
> +		esac
> +		c=$((++c))
> +	done

What happens with `git grep -e a -e b`?  Do we trigger into ref
completion too early when we should still be doing the regex
completion?

-- 
Shawn.

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [RFC PATCH] bash completion: complete refs for git-grep
  2009-10-06 15:45 ` Shawn O. Pearce
@ 2009-10-07 15:27   ` Thomas Rast
  2009-10-12  9:00     ` [PATCH] " Thomas Rast
  0 siblings, 1 reply; 6+ messages in thread
From: Thomas Rast @ 2009-10-07 15:27 UTC (permalink / raw)
  To: Shawn O. Pearce; +Cc: git

Shawn O. Pearce wrote:
> 
> Thomas Rast <trast@student.ethz.ch> wrote:
> > +	local i c=1 have_regex=""
> > +	while [ $c -lt $COMP_CWORD ]; do
> > +		i="${COMP_WORDS[c]}"
> > +		case "$i" in
> > +		-e) ;;
> > +		-e*) have_regex="$c" ; break ;;
> > +		-*) ;;
> > +		*) have_regex="$c"; break ;;
> > +		esac
> > +		c=$((++c))
> > +	done
> 
> What happens with `git grep -e a -e b`?  Do we trigger into ref
> completion too early when we should still be doing the regex
> completion?

Hmm, true, I would also have to check for the last argument (before
completion) being -e.

However, that is kind of moot because we currently complete filenames
anyway, and you said I can't stop that:

> > This is still RFC because, as you can see in the code below, I tried
> > to avoid completing at all while the user still needs to supply a
> > regex.  Sadly, bash turns the COMPREPLY=() into filename completion
> > anyway.  Is there a way to prevent this?
> 
> Not that I know of.  You can turn off default filename completion
> when you register the completion function, but that then breaks
> like every other git command for completion support because a lot
> of them do want to complete filenames.

So I'll roll a simpler patch that just always (before --) completes
refs instead, if that's ok.

-- 
Thomas Rast
trast@{inf,student}.ethz.ch

^ permalink raw reply	[flat|nested] 6+ messages in thread

* [PATCH] bash completion: complete refs for git-grep
  2009-10-07 15:27   ` Thomas Rast
@ 2009-10-12  9:00     ` Thomas Rast
  2009-10-12 14:27       ` Shawn O. Pearce
  0 siblings, 1 reply; 6+ messages in thread
From: Thomas Rast @ 2009-10-12  9:00 UTC (permalink / raw)
  To: Shawn O. Pearce; +Cc: git

Before the --, always attempt ref completion.  This helps with
entering the <treeish> arguments to git-grep.  As a bonus, you can
work around git-grep's current lack of --all by hitting M-*, ugly as
the resulting command line may be.

Strictly speaking, completing the regular expression argument (or
option argument) makes no sense.  However, we cannot prevent _all_
completion (it will fall back to filenames), so we dispense with any
additional complication to detect whether the user still has to enter
a regular expression.

Signed-off-by: Thomas Rast <trast@student.ethz.ch>
---

Sorry for taking so long; I was swamped all weekend by, well, the
weekend. ;-)

I wrote:
> Shawn O. Pearce wrote:
> > > This is still RFC because, as you can see in the code below, I tried
> > > to avoid completing at all while the user still needs to supply a
> > > regex.  Sadly, bash turns the COMPREPLY=() into filename completion
> > > anyway.  Is there a way to prevent this?
> > 
> > Not that I know of.  You can turn off default filename completion
> > when you register the completion function, but that then breaks
> > like every other git command for completion support because a lot
> > of them do want to complete filenames.
> 
> So I'll roll a simpler patch that just always (before --) completes
> refs instead, if that's ok.

So that's what this patch does.


 contrib/completion/git-completion.bash |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/contrib/completion/git-completion.bash b/contrib/completion/git-completion.bash
index 6fd7e1d..b08cd77 100755
--- a/contrib/completion/git-completion.bash
+++ b/contrib/completion/git-completion.bash
@@ -1048,7 +1048,8 @@ _git_grep ()
 		return
 		;;
 	esac
-	COMPREPLY=()
+
+	__gitcomp "$(__git_refs)"
 }
 
 _git_help ()
-- 
1.6.5.61.g35405

^ permalink raw reply related	[flat|nested] 6+ messages in thread

* Re: [PATCH] bash completion: complete refs for git-grep
  2009-10-12  9:00     ` [PATCH] " Thomas Rast
@ 2009-10-12 14:27       ` Shawn O. Pearce
  2009-10-12 23:42         ` Junio C Hamano
  0 siblings, 1 reply; 6+ messages in thread
From: Shawn O. Pearce @ 2009-10-12 14:27 UTC (permalink / raw)
  To: Thomas Rast; +Cc: git

Thomas Rast <trast@student.ethz.ch> wrote:
> > So I'll roll a simpler patch that just always (before --) completes
> > refs instead, if that's ok.

Hard to argue with that logic.  :-)

Acked-by: Shawn O. Pearce <spearce@spearce.org>

> diff --git a/contrib/completion/git-completion.bash b/contrib/completion/git-completion.bash
> index 6fd7e1d..b08cd77 100755
> --- a/contrib/completion/git-completion.bash
> +++ b/contrib/completion/git-completion.bash
> @@ -1048,7 +1048,8 @@ _git_grep ()
>  		return
>  		;;
>  	esac
> -	COMPREPLY=()
> +
> +	__gitcomp "$(__git_refs)"
>  }

-- 
Shawn.

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] bash completion: complete refs for git-grep
  2009-10-12 14:27       ` Shawn O. Pearce
@ 2009-10-12 23:42         ` Junio C Hamano
  0 siblings, 0 replies; 6+ messages in thread
From: Junio C Hamano @ 2009-10-12 23:42 UTC (permalink / raw)
  To: Shawn O. Pearce; +Cc: Thomas Rast, git

"Shawn O. Pearce" <spearce@spearce.org> writes:

> Thomas Rast <trast@student.ethz.ch> wrote:
>> > So I'll roll a simpler patch that just always (before --) completes
>> > refs instead, if that's ok.
>
> Hard to argue with that logic.  :-)
>
> Acked-by: Shawn O. Pearce <spearce@spearce.org>

Thanks.

>
>> diff --git a/contrib/completion/git-completion.bash b/contrib/completion/git-completion.bash
>> index 6fd7e1d..b08cd77 100755
>> --- a/contrib/completion/git-completion.bash
>> +++ b/contrib/completion/git-completion.bash
>> @@ -1048,7 +1048,8 @@ _git_grep ()
>>  		return
>>  		;;
>>  	esac
>> -	COMPREPLY=()
>> +
>> +	__gitcomp "$(__git_refs)"
>>  }
>
> -- 
> Shawn.
> --
> To unsubscribe from this list: send the line "unsubscribe git" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2009-10-12 23:43 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2009-10-06 10:08 [RFC PATCH] bash completion: complete refs for git-grep Thomas Rast
2009-10-06 15:45 ` Shawn O. Pearce
2009-10-07 15:27   ` Thomas Rast
2009-10-12  9:00     ` [PATCH] " Thomas Rast
2009-10-12 14:27       ` Shawn O. Pearce
2009-10-12 23:42         ` Junio C Hamano

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).