Duy Nguyen <pclouds@xxxxxxxxx> writes:
> Subject: [PATCH] Preserve errno in case case before calling sth_errno()
>
> All these locations do something like this
>
> sth_errno(..., somefunc(...));
>
> where somefunc() can potentially change errno, which will be read by
> sth_errno(). Call somefunc separately with errno preserved to avoid
> this.
>
> Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@xxxxxxxxx>
> ---
> builtin/am.c | 8 +++++++-
> builtin/commit.c | 8 ++++++--
> builtin/init-db.c | 9 ++++++---
> rerere.c | 9 ++++++---
> shallow.c | 27 ++++++++++++++++++---------
> 5 files changed, 43 insertions(+), 18 deletions(-)
> diff --git a/builtin/init-db.c b/builtin/init-db.c
> index c9b7946bad..e384749f73 100644
> --- a/builtin/init-db.c
> +++ b/builtin/init-db.c
> @@ -570,9 +570,12 @@ int cmd_init_db(int argc, const char **argv, const char *prefix)
> set_git_work_tree(work_tree);
> else
> set_git_work_tree(git_work_tree_cfg);
> - if (access(get_git_work_tree(), X_OK))
> - die_errno (_("Cannot access work tree '%s'"),
> - get_git_work_tree());
> + if (access(get_git_work_tree(), X_OK)) {
> + int saved_errno = errno;
> + const char *path = get_git_work_tree();
> + errno = saved_errno;
> + die_errno(_("Cannot access work tree '%s'"), path);
> + }
> }
This one is the most faithful conversion from "mechanical rewrite"
point of view, but I wonder if we should instead take the returned
path from get_git_work_tree() and use it in both calls. After all,
this is hardly performance sensitive codepath, so even "an obviously
safe but wasteful with extra xstrdup/free" version
work_tree_path = xstrdup(get_git_work_tree());
if (access(work_tree_path, X_OK))
die_errno(_("msg..."), work_tree_path);
free(work_tree_path);
may be an improvement.
> diff --git a/rerere.c b/rerere.c
> index 1ce440f4bb..f19a53ff2c 100644
> --- a/rerere.c
> +++ b/rerere.c
> @@ -683,9 +683,12 @@ static int merge(const struct rerere_id *id, const char *path)
> * A successful replay of recorded resolution.
> * Mark that "postimage" was used to help gc.
> */
> - if (utime(rerere_path(id, "postimage"), NULL) < 0)
> - warning_errno("failed utime() on %s",
> - rerere_path(id, "postimage"));
> + if (utime(rerere_path(id, "postimage"), NULL) < 0) {
> + int saved_errno = errno;
> + const char *path = rerere_path(id, "postimage");
> + errno = saved_errno;
> + warning_errno("failed utime() on %s", path);
> + }
Likewise.
> diff --git a/shallow.c b/shallow.c
> index df4d44ea7a..9da82f5292 100644
> --- a/shallow.c
> +++ b/shallow.c
> @@ -295,9 +295,12 @@ const char *setup_temporary_shallow(const struct oid_array *extra)
> temp = xmks_tempfile(git_path("shallow_XXXXXX"));
>
> if (write_in_full(temp->fd, sb.buf, sb.len) < 0 ||
> - close_tempfile_gently(temp) < 0)
> - die_errno("failed to write to %s",
> - get_tempfile_path(temp));
> + close_tempfile_gently(temp) < 0) {
> + int saved_errno = errno;
> + const char *path = get_tempfile_path(temp);
> + errno = saved_errno;
> + die_errno("failed to write to %s", path);
> + }
It feels a bit questionable to my taste to pretend that we are truly
oblivious to how trivial get_tempfile_path() is, i.e. no more than
just a few field accesses to "tempfile" struct. It buries more
important thing that is happening in the code in noise.
> @@ -319,9 +322,12 @@ void setup_alternate_shallow(struct lock_file *shallow_lock,
Likewise.
> @@ -366,9 +372,12 @@ void prune_shallow(int show_only)
Likewise.
All others I snipped looked like good changes. Thanks.
