Re: [PATCH v5 4/4] completion: fix zsh alias listing for subsection aliases
- From
D. Ben Knoble <ben.knoble@gmail.com>
- Date
- Feb 16, 2026, 18:32 UTC
- Message-ID
- <CALnO6CD3QpEZ=689w3mQpc5Hx0bj3+E8gU-JhstYJ9XOvunxiQ@mail.gmail.com>
- In-Reply-To
- <20260216161513.2533141-5-jonatan@jontes.page>
On Mon, Feb 16, 2026 at 11:15 AM Jonatan Holmgren <jonatan@jontes.page> wrote:
Show 18 quoted lines
> > The zsh completion function __git_zsh_cmd_alias() uses 'git config > --get-regexp' to enumerate aliases and then strips the "alias." prefix > from each key. For subsection-style aliases (alias.name.command), this > leaves "name.command" as the completion candidate instead of just > "name". > > The bash completion does not have this problem because it goes through > 'git --list-cmds=alias', which calls list_aliases() in C and already > handles both alias syntaxes correctly. However, zsh needs both the > alias name and its value for descriptive completion, which > --list-cmds=alias does not provide. > > Add a hidden --aliases-for-completion option to 'git help', following > the existing --config-for-completion pattern. It outputs NUL-separated > "name\nvalue" pairs using list_aliases(), which correctly resolves both > the traditional (alias.name) and subsection (alias.name.command) > formats. Update __git_zsh_cmd_alias() to use it.
An alternative would to be to post-process and turn "alias.name.command" into "name", but this solution is easier to parse (won't mix 2 formats).
Show 34 quoted lines
> ---
> builtin/help.c | 15 +++++++++++++++
> contrib/completion/git-completion.zsh | 2 +-
> 2 files changed, 16 insertions(+), 1 deletion(-)
>
> diff --git a/builtin/help.c b/builtin/help.c
> index c09cbc8912..f02308a391 100644
> --- a/builtin/help.c
> +++ b/builtin/help.c
> @@ -54,6 +54,7 @@ static enum help_action {
> HELP_ACTION_DEVELOPER_INTERFACES,
> HELP_ACTION_CONFIG_FOR_COMPLETION,
> HELP_ACTION_CONFIG_SECTIONS_FOR_COMPLETION,
> + HELP_ACTION_ALIASES_FOR_COMPLETION,
> } cmd_mode;
>
> static char *html_path;
> @@ -90,6 +91,8 @@ static struct option builtin_help_options[] = {
> HELP_ACTION_CONFIG_FOR_COMPLETION, PARSE_OPT_HIDDEN),
> OPT_CMDMODE_F(0, "config-sections-for-completion", &cmd_mode, "",
> HELP_ACTION_CONFIG_SECTIONS_FOR_COMPLETION, PARSE_OPT_HIDDEN),
> + OPT_CMDMODE_F(0, "aliases-for-completion", &cmd_mode, "",
> + HELP_ACTION_ALIASES_FOR_COMPLETION, PARSE_OPT_HIDDEN),
>
> OPT_END(),
> };
> @@ -691,6 +694,18 @@ int cmd_help(int argc,
> help_format);
> list_config_help(SHOW_CONFIG_SECTIONS);
> return 0;
> + case HELP_ACTION_ALIASES_FOR_COMPLETION: {
> + struct string_list alias_list = STRING_LIST_INIT_DUP;
> + opt_mode_usage(argc, "--aliases-for-completion",
> + help_format);style nit: I _think_ this line could be unwrapped?
> + list_aliases(&alias_list); > + string_list_sort(&alias_list);
It may not matter much since lists of aliases are probably not humongous, but do we need to sort? I don't _think_ "git --list-cmds=alias" does, for example, though I'm not sure about "git config --get-regexp" that's being replaced. Leaving it in the config order could even be a feature, so I can put easier-to-complete aliases earlier in my config?
Show 9 quoted lines
> + for (size_t i = 0; i < alias_list.nr; i++)
> + printf("%s%c%s%c", alias_list.items[i].string, '\n',
> + (char *)alias_list.items[i].util, '\0');
> + string_list_clear(&alias_list, 1);
> + return 0;
> + }
> case HELP_ACTION_CONFIG:
> opt_mode_usage(argc, "--config", help_format);
> setup_pager(the_repository);Show 15 quoted lines
> diff --git a/contrib/completion/git-completion.zsh b/contrib/completion/git-completion.zsh
> index f5877bd7a1..c32186a977 100644
> --- a/contrib/completion/git-completion.zsh
> +++ b/contrib/completion/git-completion.zsh
> @@ -202,7 +202,7 @@ __git_zsh_cmd_common ()
> __git_zsh_cmd_alias ()
> {
> local -a list
> - list=(${${(0)"$(git config -z --get-regexp '^alias\.*')"}#alias.})
> + list=(${(0)"$(git help --aliases-for-completion)"})
> list=(${(f)"$(printf "%s:alias for '%s'\n" ${(f@)list})"})
> _describe -t alias-commands 'aliases' list && _ret=0
> }
> --
> 2.53.0.83.g660bbd62ee.dirtyThe Zsh hunk looks straightforwardly correct, especially since the output format of the new help-mode is unchanged, thanks.
-- D. Ben Knoble