From: Daniel Shahaf <d.s@daniel.shahaf.name>
To: "Jun T." <takimoto-j@kba.biglobe.ne.jp>
Cc: zsh-workers@zsh.org
Subject: Re: _git-reset doesn't complete newly staged files
Date: Wed, 09 Mar 2016 13:24:44 +0000 [thread overview]
Message-ID: <20160309132444.GA2428@tarsus.local2> (raw)
In-Reply-To: <C2B8CB27-F1B7-48C5-B343-A7AA86049CD8@kba.biglobe.ne.jp> <234B98BE-D775-4068-AE44-81715F430E15@kba.biglobe.ne.jp>
Jun T. wrote on Mon, Mar 07, 2016 at 12:16:54 +0900:
> Are there any comments to my patch? Especially from those
> who know git better.
> If there is nothing obviously wrong, I'll push it.
Sorry for not replying earlier: I had this flagged for attention but
haven't spoken up sooner since I expected I wouldn't have time to reply
to responses.
More below.
Jun T. wrote on Thu, Mar 03, 2016 at 14:07:12 +0900:
> In the case of
>
> % git reset HEAD <TAB>
>
> the patch below tries to offer only staged files (including
> newly staged files which have never been committed).
> A new function __git_staged_files() ignores unmerged (conflicting)
> files, but I'm not sure this is correct or not.
>
> In the case of
>
> % git reset commit <TAB> # commit is not HEAD
>
> I have no idea what to do, so I just added __git_ignore_line
> to the original code.
>
>
>
> diff --git a/Completion/Unix/Command/_git b/Completion/Unix/Command/_git
> index b7eaf2e..7a459f1 100644
> --- a/Completion/Unix/Command/_git
> +++ b/Completion/Unix/Command/_git
> @@ -1377,7 +1377,11 @@ _git-reset () {
> if [[ -n $line[1] ]] && __git_is_committish $line[1]; then
> commit=$line[1]
> fi
This should actually call __git_is_treeish, not __git_is_committish.
(That's a preëxisting problem: it precedes your patch.)
> - __git_tree_files ${PREFIX:-.} $commit && ret=0
> + if [[ $commit == HEAD ]]; then
As you say, special-casing HEAD this way is not wrong, but shouldn't be
needed: "HEAD" is a treeish and should be treated like any other
treeish.
I think you're looking for «git diff-index --cached -z --name-only
$treeish». That gives the names of files that differ between the index
and $treeish. (I wrote «$treeish» in pseudocode, but the real code's
local parameter «$commit» contains exactly the right value.)
I'm not sure how to handle the syntax 'git reset foo' where 'foo' is
both a file in the index and a treeish (e.g., a tag name).
Incidentally, it would be nice to have a docstring for __git_ignore_line:
diff --git a/Completion/Unix/Command/_git b/Completion/Unix/Command/_git
index b7eaf2e..8043bb9 100644
--- a/Completion/Unix/Command/_git
+++ b/Completion/Unix/Command/_git
@@ -5034,6 +5034,16 @@ __git_describe_commit () {
# Completion Wrappers
+# '__git_ignore_line $callee "${callee_args[@]}" "${callee_compadd_args[@]}"'
+# invokes '$callee "${callee_args[@]}" "${callee_compadd_args[@]}"' with
+# callee_compadd_args modified to exclude positional parameters to the completed
+# command from being completed. This causes 'git add foo <TAB>' not to offer
+# 'foo' again.
+#
+# Note: This function can't be used to wrap bare 'compadd' calls that use a '--'
+# argument terminator. It can wrap functions of the form
+# f() { shift $N; compadd "$@" -a - mymatches }
+# .
(( $+functions[__git_ignore_line] )) ||
__git_ignore_line () {
declare -a ignored
Why does __git_ignore_line escape some characters with backslashes
before adding them to $ignored?
Thanks,
Daniel
> + __git_ignore_line __git_staged_files && ret=0
> + else
> + __git_ignore_line __git_tree_files ${PREFIX:-.} $commit && ret=0
> + fi
> ;;
> esac
>
> @@ -6131,6 +6135,29 @@ __git_tree_files () {
> _wanted files expl 'tree file' _multi_parts -f $compadd_opts -- / tree_files
> }
>
> +(( $+functions[__git_staged_files] )) ||
> +__git_staged_files () {
> + local -a slist staged_files
> + local item expl i
> +
> + slist=(${(0)"$(_call_program staged-files git status -z -uno 2>/dev/null)"})
> + __git_command_successful $pipestatus || return 1
> +
> + for (( i = 1; i <= $#slist; ++i )) do
> + item=$slist[i]
> + if [[ $item == (DD|AA|U|?U)* ]]; then
> + continue #XXX skip unmerged files
> + elif [[ $item == R* ]]; then
> + staged_files+=( $item[4,-1] $slist[++i] )
> + elif [[ $item == [ACDM]* ]]; then
> + staged_files+=( $item[4,-1] )
> + fi
> + done
> + staged_files=( ${(0)"$(__git_files_relative ${(pj:\0:)staged_files})"} )
> +
> + _wanted staged-files expl 'staged file' compadd "$@" -a - staged_files
> +}
> +
> # Repository Argument Types
>
> (( $+functions[__git_remote_repositories] )) ||
>
>
>
>
>
next prev parent reply other threads:[~2016-03-09 13:31 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-02-26 8:08 Jun T.
2016-03-03 5:07 ` Jun T.
2016-03-07 3:16 ` Jun T.
2016-03-09 13:24 ` Daniel Shahaf [this message]
2016-03-09 23:33 ` Daniel Shahaf
2016-03-10 23:15 ` [PATCH] _git: Fix completion of diffs against the index when treeish isn't shell-safe Daniel Shahaf
2016-03-10 12:03 ` _git-reset doesn't complete newly staged files Jun T.
2016-03-10 14:35 ` Mikael Magnusson
2016-03-10 23:20 ` __git_ignore_line positional argument (un)escaping (was: Re: _git-reset doesn't complete newly staged files) Daniel Shahaf
2016-03-11 8:24 ` Jun T.
2016-03-11 22:13 ` Daniel Shahaf
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20160309132444.GA2428@tarsus.local2 \
--to=d.s@daniel.shahaf.name \
--cc=takimoto-j@kba.biglobe.ne.jp \
--cc=zsh-workers@zsh.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://git.vuxu.org/mirror/zsh/
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).