Small patches o git completion

Small patches o git completion

[Nikolai Weibull]

This patch series fixes some small issues raised on the mailing list in regard
to the git completion.

[PATCH 1/4] Fix typo

[Nikolai Weibull]

Signed-off-by: Nikolai Weibull <now@bitwi.se>
---
 Completion/Unix/Command/_git |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/Completion/Unix/Command/_git b/Completion/Unix/Command/_git
index bf22a3c..3694d4a 100644
--- a/Completion/Unix/Command/_git
+++ b/Completion/Unix/Command/_git
@@ -3070,7 +3070,7 @@ _git-cvsexportcommit () {
     '-m[prepend the commit message with the provided prefix]:message prefix' \
     '-u[update affected files from CVS repository before attempting export]' \
     '-k[reverse CVS keyword expansion]' \
-    '-w[specify location of cVS checkout to use for export]' \
+    '-w[specify location of CVS checkout to use for export]' \
     '-W[use current working directory for bot Git and CVS checkout]' \
     '-v[verbose output]' \
     '-h[display usage]' \
-- 
1.7.4.1

[PATCH 2/4] Only show modified files for git-checkout without tree

[Nikolai Weibull]

Signed-off-by: Nikolai Weibull <now@bitwi.se>
---
 Completion/Unix/Command/_git |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/Completion/Unix/Command/_git b/Completion/Unix/Command/_git
index 3694d4a..8a590be 100644
--- a/Completion/Unix/Command/_git
+++ b/Completion/Unix/Command/_git
@@ -396,7 +396,7 @@ _git-checkout () {
         # TODO: Allow A...B
         local branch_arg='branches::__git_revisions' \
               tree_ish_arg='tree-ishs::__git_tree_ishs' \
-              file_arg='files::__git_cached_files'
+              file_arg='modified-files::__git_modified_files'
 
         if [[ -n ${opt_args[(I)-b|-B|--orphan]} ]]; then
           tree_ish_arg=
@@ -418,7 +418,7 @@ _git-checkout () {
       elif [[ -n $line[1] ]] && __git_is_treeish $line[1]; then
         __git_ignore_line __git_tree_files ${PREFIX:-.} $line[1] && ret=0
       else
-        __git_ignore_line __git_cached_files && ret=0
+        __git_ignore_line __git_modified_files && ret=0
       fi
       ;;
   esac
-- 
1.7.4.1

[PATCH 3/4] Fix git-branch -[dD] -r completion

[Nikolai Weibull]

Signed-off-by: Nikolai Weibull <now@bitwi.se>
---
 Completion/Unix/Command/_git |   10 +++++++---
 1 files changed, 7 insertions(+), 3 deletions(-)

diff --git a/Completion/Unix/Command/_git b/Completion/Unix/Command/_git
index 8a590be..0f573a1 100644
--- a/Completion/Unix/Command/_git
+++ b/Completion/Unix/Command/_git
@@ -266,8 +266,12 @@ _git-branch () {
   if (( words[(I)-d] || words[(I)-D] )); then
     dependent_creation_args=
     dependent_deletion_args=(
-      '-r[delete remote-tracking branches]'
-      '*: :__git_ignore_line_inside_arguments __git_branch_names')
+      '-r[delete remote-tracking branches]')
+    if (( words[(I)-r] )); then
+      dependent_deletion_args+='*: :__git_ignore_line_inside_arguments __git_remote_branch_names'
+    else
+      dependent_deletion_args+='*: :__git_ignore_line_inside_arguments __git_branch_names'
+    fi
   fi
 
   declare -a dependent_modification_args
@@ -4913,7 +4917,7 @@ __git_remote_branch_names () {
   local expl
   declare -a branch_names
 
-  branch_names=(${${(f)"$(_call_program remote-branch-refs git for-each-ref --format='%(refname)' refs/remotes 2>/dev/null)"}#refs/remotes/})
+  branch_names=(${${(f)"$(_call_program remote-branch-refs git for-each-ref --format='"%(refname)"' refs/remotes 2>/dev/null)"}#refs/remotes/})
   __git_command_successful $pipestatus || return
 
   _wanted remote-branch-names expl 'remote branch name' compadd $* - $branch_names
-- 
1.7.4.1

[PATCH 4/4] Fix typo in git-send-email completion

[Nikolai Weibull]

Signed-off-by: Nikolai Weibull <now@bitwi.se>
---
 Completion/Unix/Command/_git |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/Completion/Unix/Command/_git b/Completion/Unix/Command/_git
index 0f573a1..b4b36f4 100644
--- a/Completion/Unix/Command/_git
+++ b/Completion/Unix/Command/_git
@@ -3161,7 +3161,7 @@ _git-send-email () {
     '--smtp-pass=[specify password to use for SMTP-AUTH]::password' \
     '--smtp-server=[specify SMTP server to connect to]:smtp server:_hosts' \
     '--smtp-server-port=[specify port to connect to SMTP server on]:smtp port:_ports' \
-    '--smtp-user=[specify user to use for SMTP-AUTH:smtp user:_users' \
+    '--smtp-user=[specify user to use for SMTP-AUTH]:smtp user:_users' \
     '--cc-cmd=[specify command to generate Cc\: header with]:Cc\: command:_path_commands' \
     '(                 --no-chain-reply-to)--chain-reply-to[send each email as a reply to previous one]' \
     '(--chain-reply-to                    )--no-chain-reply-to[send all emails after first as replies to first one]' \
-- 
1.7.4.1

Re: [PATCH 3/4] Fix git-branch -[dD] -r completion

[Frank Terbeck]

Hm, with this patch, this:

% git branch -d -r <tab>

works. But this:

% git -r -d <tab>

doesn't.

I suppose, that's not intentional?

Regards, Frank

-- 
In protocol design, perfection has been reached not when there is
nothing left to add, but when there is nothing left to take away.
                                                  -- RFC 1925

Re: Small patches o git completion

[Frank Terbeck]

Nikolai Weibull wrote:
> This patch series fixes some small issues raised on the mailing list
> in regard to the git completion.

I've committed patches 1, 2 and 4 of this series. See my reply to patch
3/4 as to why I left it out for now.

Regards, Frank

Re: [PATCH 3/4] Fix git-branch -[dD] -r completion

[Benjamin R. Haskell]

On Fri, 18 Mar 2011, Frank Terbeck wrote:

> Hm, with this patch, this:
>
> % git branch -d -r <tab>
>
> works. But this:
>
> % git -r -d <tab>
>
> doesn't.

Without 'branch'?


> I suppose, that's not intentional?

With the 'branch' there, both:

(i) git branch -d -r <tab>
and
(ii) git branch -r -d <tab>

seem to behave identically whether or not this patch is applied.  With 
or without the patch (i) seems fine AFAICT.  But (ii) doesn't seem to 
complete branches, even though I think it should.

-- 
Best,
Ben

Re: [PATCH 3/4] Fix git-branch -[dD] -r completion

[Frank Terbeck]

Benjamin R. Haskell wrote:
> On Fri, 18 Mar 2011, Frank Terbeck wrote:
>
>> Hm, with this patch, this:
>>
>> % git branch -d -r <tab>
>>
>> works. But this:
>>
>> % git -r -d <tab>
>>
>> doesn't.
>
> Without 'branch'?

Sorry typo. It doesn't work *with* branch.

>> I suppose, that's not intentional?
>
> With the 'branch' there, both:
>
> (i) git branch -d -r <tab>
> and
> (ii) git branch -r -d <tab>
>
> seem to behave identically whether or not this patch is applied.  With
> or without the patch (i) seems fine AFAICT.  But (ii) doesn't seem to
> complete branches, even though I think it should.

They don't for me.

% git branch -d -r <tab>
- remote branch name -
origin/debian           origin/upstream       
origin/HEAD             origin/upstream-import
origin/master           zsh.git/master

And:

% git branch -r -d <tab>

results in

% git branch -r -d -

If I tab again, I get:

% git branch -r -d -<tab>
- option -
--abbrev         -- set minimum SHA1 display-length                                                                                                                 
--color          -- turn on branch coloring                                                                                                                         
--no-abbrev      -- do not abbreviate sha1s                                                                                                                         
--no-color       -- turn off branch coloring                                                                                                                        
-r               -- delete remote-tracking branches                                                                                                                 
--verbose    -v  -- show SHA1 and commit subject line for each head

Regards, Frank

-- 
In protocol design, perfection has been reached not when there is
nothing left to add, but when there is nothing left to take away.
                                                  -- RFC 1925

Re: [PATCH 3/4] Fix git-branch -[dD] -r completion

[Nikolai Weibull]

On Fri, Mar 18, 2011 at 23:26, Frank Terbeck <ft@bewatermyfriend.org> wrote:
> Hm, with this patch, this:
>
> % git branch -d -r <tab>
>
> works. But this:
>
> % git -r -d <tab>
>
> doesn't.
>
> I suppose, that's not intentional?

Well, it depends on what you’re after.  Writing -r first hints that
you want to list branches, not delete them.  If you look at the
synopsis in the man page, that’s how I’ve done it.  Does anyone have a
strong opinion on how this should be done?

Re: [PATCH 3/4] Fix git-branch -[dD] -r completion

[Frank Terbeck]

Nikolai Weibull wrote:
> On Fri, Mar 18, 2011 at 23:26, Frank Terbeck <ft@bewatermyfriend.org> wrote:
>> Hm, with this patch, this:
>>
>> % git branch -d -r <tab>
>>
>> works. But this:
>>
>> % git -r -d <tab>
>>
>> doesn't.
>>
>> I suppose, that's not intentional?
>
> Well, it depends on what you’re after.  Writing -r first hints that
> you want to list branches, not delete them.  If you look at the
> synopsis in the man page, that’s how I’ve done it.  Does anyone have a
> strong opinion on how this should be done?

I think _git should offer both ways simply because git accepts both
ways. Further down the git-branch(1) manpage, it states "List or delete
(if used with -d) [...]". So it's *with*, not *after*.

IMHO, _git behaving that way would make sense.

Regards, Frank

Re: [PATCH 3/4] Fix git-branch -[dD] -r completion

[Benjamin R. Haskell]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 655 bytes --]

On Sat, 19 Mar 2011, Nikolai Weibull wrote:

> On Fri, Mar 18, 2011 at 23:26, Frank Terbeck wrote:
>> Hm, with this patch, this:
>>
>> % git branch -d -r <tab>
>>
>> works. But this:
>>
>> % git -r -d <tab>
>>
>> doesn't.
>>
>> I suppose, that's not intentional?
>
> Well, it depends on what you’re after.  Writing -r first hints that 
> you want to list branches, not delete them.  If you look at the 
> synopsis in the man page, that’s how I’ve done it.  Does anyone have a 
> strong opinion on how this should be done?

It should complete remote branches in both cases, since git accepts 
remote branch(es) to delete in both cases.

-- 
Best,
Ben

Re: Small patches o git completion

[Simon Ruderich]

[-- Attachment #1: Type: text/plain, Size: 993 bytes --]

On Fri, Mar 18, 2011 at 10:35:08PM +0100, Nikolai Weibull wrote:
> This patch series fixes some small issues raised on the mailing
> list in regard to the git completion.

Hi,

I found another minor issue:

    git tag -v <Tab>

Does nothing. Without -v it works fine.

This patch seems to fix it though I don't really understand what
it does (copied from -d):

--- a/Completion/Unix/Command/_git
+++ b/Completion/Unix/Command/_git
@@ -1511,7 +1511,9 @@ _git-tag () {
       '-l[list tags matching pattern]:pattern' \
       '--contains=[only list tags which contain the specified commit]: :__git_commits' \
     - verification \
-      '-v[verifies gpg signutare of tags]' && ret=0
+      '-v[verifies gpg signutare of tags]' \
+      '*:: :__git_ignore_line_inside_arguments __git_tags' \
+      && ret=0
 }

 # Ancillary Commands (Manipulators)

Regards,
Simon
-- 
+ privacy is necessary
+ using gnupg http://gnupg.org
+ public key id: 0x92FEFDB7E44C32F9

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]

Re: Small patches o git completion

[Simon Ruderich]

[-- Attachment #1: Type: text/plain, Size: 479 bytes --]

On Sat, Mar 19, 2011 at 11:15:48PM +0100, Simon Ruderich wrote:
> Hi,
>
> I found another minor issue:
>
>     git tag -v <Tab>
>
> Does nothing. Without -v it works fine.
>
> This patch seems to fix it though I don't really understand what
> it does (copied from -d):
>
> [snip]

Hi,

Could somebody please apply this patch. Works fine for me.

Regards,
Simon
-- 
+ privacy is necessary
+ using gnupg http://gnupg.org
+ public key id: 0x92FEFDB7E44C32F9

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]

Re: Small patches o git completion

[Nikolai Weibull]

On Fri, Apr 15, 2011 at 16:08, Simon Ruderich <simon@ruderich.org> wrote:
> On Sat, Mar 19, 2011 at 11:15:48PM +0100, Simon Ruderich wrote:
>> Hi,
>>
>> I found another minor issue:
>>
>>     git tag -v <Tab>
>>
>> Does nothing. Without -v it works fine.
>>
>> This patch seems to fix it though I don't really understand what
>> it does (copied from -d):

> Could somebody please apply this patch. Works fine for me.

Seconded.

Re: Small patches o git completion

[Mikael Magnusson]

On 15 April 2011 16:24, Nikolai Weibull <now@bitwi.se> wrote:
> On Fri, Apr 15, 2011 at 16:08, Simon Ruderich <simon@ruderich.org> wrote:
>> On Sat, Mar 19, 2011 at 11:15:48PM +0100, Simon Ruderich wrote:
>>> Hi,
>>>
>>> I found another minor issue:
>>>
>>>     git tag -v <Tab>
>>>
>>> Does nothing. Without -v it works fine.
>>>
>>> This patch seems to fix it though I don't really understand what
>>> it does (copied from -d):
>
>> Could somebody please apply this patch. Works fine for me.
>
> Seconded.

How did it complete tags without the *:foo:__git_tags line without -v?

-- 
Mikael Magnusson

Re: Small patches o git completion

[Nikolai Weibull]

On Fri, Apr 15, 2011 at 16:37, Mikael Magnusson <mikachu@gmail.com> wrote:
> On 15 April 2011 16:24, Nikolai Weibull <now@bitwi.se> wrote:
>> On Fri, Apr 15, 2011 at 16:08, Simon Ruderich <simon@ruderich.org> wrote:
>>> On Sat, Mar 19, 2011 at 11:15:48PM +0100, Simon Ruderich wrote:
>>>> Hi,
>>>>
>>>> I found another minor issue:
>>>>
>>>>     git tag -v <Tab>
>>>>
>>>> Does nothing. Without -v it works fine.
>>>>
>>>> This patch seems to fix it though I don't really understand what
>>>> it does (copied from -d):
>>
>>> Could somebody please apply this patch. Works fine for me.
>>
>> Seconded.
>
> How did it complete tags without the *:foo:__git_tags line without -v?

In what context?  There are four different option groups.

Re: Small patches o git completion

[Frank Terbeck]

Simon Ruderich wrote:
> Could somebody please apply this patch. Works fine for me.

[x] done.

It's probably a good idea to ping patches like this. I for one am not
keeping track very well.

Regards, Frank

-- 
In protocol design, perfection has been reached not when there is
nothing left to add, but when there is nothing left to take away.
                                                  -- RFC 1925

Re: Small patches o git completion

[Mikael Magnusson]

On 15 April 2011 16:44, Nikolai Weibull <now@bitwi.se> wrote:
> On Fri, Apr 15, 2011 at 16:37, Mikael Magnusson <mikachu@gmail.com> wrote:
>> On 15 April 2011 16:24, Nikolai Weibull <now@bitwi.se> wrote:
>>> On Fri, Apr 15, 2011 at 16:08, Simon Ruderich <simon@ruderich.org> wrote:
>>>> On Sat, Mar 19, 2011 at 11:15:48PM +0100, Simon Ruderich wrote:
>>>>> Hi,
>>>>>
>>>>> I found another minor issue:
>>>>>
>>>>>     git tag -v <Tab>
>>>>>
>>>>> Does nothing. Without -v it works fine.
>>>>>
>>>>> This patch seems to fix it though I don't really understand what
>>>>> it does (copied from -d):
>>>
>>>> Could somebody please apply this patch. Works fine for me.
>>>
>>> Seconded.
>>
>> How did it complete tags without the *:foo:__git_tags line without -v?
>
> In what context?  There are four different option groups.

Ah, somehow i read that - as a diff -. While you're at it, -l doesn't
take an option, the pattern is just a regular argument.
       git tag [-n[<num>]] -l [--contains <commit>] [<pattern>]

-- 
Mikael Magnusson

[PATCH] Fix -l option for git-tag completion

[Nikolai Weibull]

---
 Completion/Unix/Command/_git |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/Completion/Unix/Command/_git b/Completion/Unix/Command/_git
index 4664cfa..0135cb9 100644
--- a/Completion/Unix/Command/_git
+++ b/Completion/Unix/Command/_git
@@ -1509,12 +1509,12 @@ _git-tag () {
       '*:: :__git_ignore_line_inside_arguments __git_tags' \
     - listing \
       '-n+[limit line output of annotation]: :__git_guard_number "limit"' \
-      '-l[list tags matching pattern]:pattern' \
+      '-l[list tags matching pattern]' \
       '--contains=[only list tags which contain the specified commit]: :__git_commits' \
+      '::pattern' \
     - verification \
       '-v[verifies gpg signutare of tags]' \
-      '*:: :__git_ignore_line_inside_arguments __git_tags' \
-      && ret=0
+      '*:: :__git_ignore_line_inside_arguments __git_tags' && ret=0
 }
 
 # Ancillary Commands (Manipulators)
-- 
1.7.4.4

Re: [PATCH] Fix -l option for git-tag completion

[Mikael Magnusson]

Looks like nobody else committed this, I've done so now.

On 2 May 2011 10:16, Nikolai Weibull <now@bitwi.se> wrote:
> ---
>  Completion/Unix/Command/_git |    6 +++---
>  1 files changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/Completion/Unix/Command/_git b/Completion/Unix/Command/_git
> index 4664cfa..0135cb9 100644
> --- a/Completion/Unix/Command/_git
> +++ b/Completion/Unix/Command/_git
> @@ -1509,12 +1509,12 @@ _git-tag () {
>       '*:: :__git_ignore_line_inside_arguments __git_tags' \
>     - listing \
>       '-n+[limit line output of annotation]: :__git_guard_number "limit"' \
> -      '-l[list tags matching pattern]:pattern' \
> +      '-l[list tags matching pattern]' \
>       '--contains=[only list tags which contain the specified commit]: :__git_commits' \
> +      '::pattern' \
>     - verification \
>       '-v[verifies gpg signutare of tags]' \
> -      '*:: :__git_ignore_line_inside_arguments __git_tags' \
> -      && ret=0
> +      '*:: :__git_ignore_line_inside_arguments __git_tags' && ret=0
>  }
>
>  # Ancillary Commands (Manipulators)
> --
> 1.7.4.4

Re: [PATCH 3/4] Fix git-branch -[dD] -r completion

[Nikolai Weibull]

On Sat, Mar 19, 2011 at 14:19, Benjamin R. Haskell <zsh@benizi.com> wrote:
> On Sat, 19 Mar 2011, Nikolai Weibull wrote:
>
>> On Fri, Mar 18, 2011 at 23:26, Frank Terbeck wrote:
>>>
>>> Hm, with this patch, this:
>>>
>>> % git branch -d -r <tab>
>>>
>>> works. But this:
>>>
>>> % git -r -d <tab>
>>>
>>> doesn't.
>>>
>>> I suppose, that's not intentional?
>>
>> Well, it depends on what you’re after.  Writing -r first hints that you
>> want to list branches, not delete them.  If you look at the synopsis in the
>> man page, that’s how I’ve done it.  Does anyone have a strong opinion on how
>> this should be done?
>
> It should complete remote branches in both cases, since git accepts remote
> branch(es) to delete in both cases.

OK, this has been fixed in CVS.