From: Justin Tobler <jltobler@gmail.com>
To: Patrick Steinhardt <ps@pks.im>
Cc: git@vger.kernel.org
Subject: Re: [PATCH 1/7] builtin/init: stop modifying global `git_work_tree_cfg` variable
Date: Wed, 10 Jun 2026 16:15:01 -0500 [thread overview]
Message-ID: <ainKs-8_0w_2HhOf@denethor> (raw)
In-Reply-To: <20260610-b4-pks-setup-drop-global-state-v1-1-5dff3eec8f06@pks.im>
On 26/06/10 08:56AM, Patrick Steinhardt wrote:
> When executing git-init(1) we need to figure out the final location of
> the worktree. This location can be configured in a couple of ways: via
> an environment variable, via the preexisting "core.worktree" config in
> case we're reinitializing, or implicitly when reinitializing a non-bare
> repository.
>
> When checking for the worktree location in "builtin/init-db.c" we
> populate any potentially-discovered value both by setting the global
> `git_work_tree_cfg` variable and via `set_git_work_tree()`, which
> ultimately ends up modifying `struct repository::worktree`.
Does `git_work_tree_cfg` have any bearing on the worktree value that
ends up in `struct repository`? Or is it just `set_git_work_tree()`?
> Modifying `git_work_tree_cfg` is unnecessary though: we configure the
> worktree in `create_default_files()`, and that function derives the
> worktree location via `repo_get_work_tree()`. Consequently, propagating
> the worktree via `set_git_work_tree()` is sufficient.
Ok, so IIUC there is really no reason to globally store the
`git_work_tree_cfg` value derived during `cmd_init_db()`. We only care
about the worktree written to `struct repository`.
> Stop munging `git_work_tree_cfg` and make it file-local to "setup.c" and
> function-local to `cmd_init_db()`.
Makes sense.
> Signed-off-by: Patrick Steinhardt <ps@pks.im>
> ---
> builtin/init-db.c | 4 ++++
> environment.c | 3 ---
> environment.h | 1 -
> setup.c | 3 +++
> 4 files changed, 7 insertions(+), 4 deletions(-)
>
> diff --git a/builtin/init-db.c b/builtin/init-db.c
> index c55517ad94..01bc27904e 100644
> --- a/builtin/init-db.c
> +++ b/builtin/init-db.c
> @@ -229,6 +229,8 @@ int cmd_init_db(int argc,
>
> if (!is_bare_repository_cfg) {
> const char *git_dir_parent = strrchr(git_dir, '/');
> + char *git_work_tree_cfg = NULL;
We figure out the worktree locally and continue to set it via
`set_git_work_tree()` without updating a global worktree config value.
> +
> if (git_dir_parent) {
> char *rel = xstrndup(git_dir, git_dir_parent - git_dir);
> git_work_tree_cfg = real_pathdup(rel, 1);
> @@ -243,6 +245,8 @@ int cmd_init_db(int argc,
> if (access(repo_get_work_tree(the_repository), X_OK))
> die_errno (_("Cannot access work tree '%s'"),
> repo_get_work_tree(the_repository));
> +
> + free(git_work_tree_cfg);
> }
> else {
> if (real_git_dir)
> diff --git a/environment.c b/environment.c
> index fc3ed8bb1c..4e86335f25 100644
> --- a/environment.c
> +++ b/environment.c
> @@ -100,9 +100,6 @@ int auto_comment_line_char;
> bool warn_on_auto_comment_char;
> #endif /* !WITH_BREAKING_CHANGES */
>
> -/* This is set by setup_git_directory_gently() and/or git_default_config() */
> -char *git_work_tree_cfg;
> -
> /*
> * Repository-local GIT_* environment variables; see environment.h for details.
> */
> diff --git a/environment.h b/environment.h
> index ccfcf37bfb..5d6e4e6c1b 100644
> --- a/environment.h
> +++ b/environment.h
> @@ -149,7 +149,6 @@ int have_git_dir(void);
>
> extern int is_bare_repository_cfg;
> int is_bare_repository(void);
> -extern char *git_work_tree_cfg;
>
> /* Environment bits from configuration mechanism */
> extern int trust_executable_bit;
> diff --git a/setup.c b/setup.c
> index b4652651df..52228b42a1 100644
> --- a/setup.c
> +++ b/setup.c
> @@ -31,6 +31,9 @@ enum allowed_bare_repo {
> ALLOWED_BARE_REPO_ALL,
> };
>
> +/* This is set by setup_git_directory_gently() and/or git_default_config() */
> +static char *git_work_tree_cfg;
Ok, now `git_work_tree_cfg` is only referenced from "setup.c".
-Justin
next prev parent reply other threads:[~2026-06-10 21:15 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-06-10 6:56 [PATCH 0/7] setup: drop global state Patrick Steinhardt
2026-06-10 6:56 ` [PATCH 1/7] builtin/init: stop modifying global `git_work_tree_cfg` variable Patrick Steinhardt
2026-06-10 21:15 ` Justin Tobler [this message]
2026-06-10 6:56 ` [PATCH 2/7] builtin/init: simplify logic to configure worktree Patrick Steinhardt
2026-06-10 21:29 ` Justin Tobler
2026-06-10 6:56 ` [PATCH 3/7] setup: remove global `git_work_tree_cfg` variable Patrick Steinhardt
2026-06-10 21:52 ` Justin Tobler
2026-06-10 6:56 ` [PATCH 4/7] builtin/init: stop modifying `is_bare_repository_cfg` Patrick Steinhardt
2026-06-10 6:56 ` [PATCH 5/7] environment: split up concerns of `is_bare_repository_cfg` Patrick Steinhardt
2026-06-10 22:22 ` Justin Tobler
2026-06-10 6:56 ` [PATCH 6/7] environment: stop using `the_repository` in `is_bare_repository()` Patrick Steinhardt
2026-06-10 6:56 ` [PATCH 7/7] treewide: drop USE_THE_REPOSITORY_VARIABLE Patrick Steinhardt
2026-06-10 22:26 ` Justin Tobler
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=ainKs-8_0w_2HhOf@denethor \
--to=jltobler@gmail.com \
--cc=git@vger.kernel.org \
--cc=ps@pks.im \
/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.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox