Re: [PATCH v4 27/27] count-objects: report unused files in $GIT_DIR/repos/...
- From
Eric Sunshine <sunshine@sunshineco.com>
- Date
- Mar 5, 2014, 04:25 UTC
- Message-ID
- <CAPig+cToTqJ4PPX3mvt0ftC-+Ncfiakz5dd_=xymM76T01z4fQ@mail.gmail.com>
- In-Reply-To
- <1393675983-3232-28-git-send-email-pclouds@gmail.com>
On Sat, Mar 1, 2014 at 7:13 AM, Nguyễn Thái Ngọc Duy <pclouds@gmail.com> wrote:
Show 78 quoted lines
> In linked checkouts, borrowed parts like config is taken from
> $GIT_COMMON_DIR. $GIT_DIR/config is never used. Report them as
> garbage.
>
> Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
> ---
> builtin/count-objects.c | 37 ++++++++++++++++++++++++++++++++++++-
> path.c | 4 ++++
> 2 files changed, 40 insertions(+), 1 deletion(-)
>
> diff --git a/builtin/count-objects.c b/builtin/count-objects.c
> index a7f70cb..725cd5f 100644
> --- a/builtin/count-objects.c
> +++ b/builtin/count-objects.c
> @@ -78,6 +78,39 @@ static void count_objects(DIR *d, char *path, int len, int verbose,
> }
> }
>
> +static void report_linked_checkout_garbage(void)
> +{
> + /*
> + * must be more or less in sync with * path.c:update_common_dir().
> + *
> + * "logs" is let slip because logs/HEAD is in $GIT_DIR but the
> + * remaining in $GIT_COMMON_DIR. Probably not worth traversing
> + * the entire "logs" directory for that.
> + *
> + * The same "gc.pid" for because it's a temporary file.
> + */
> + const char *list[] = {
> + "branches", "hooks", "info", "lost-found", "modules",
> + "objects", "refs", "remotes", "rr-cache", "svn",
> + "config", "packed-refs", "shallow", NULL
> + };
> + struct strbuf sb = STRBUF_INIT;
> + const char **p;
> + int len;
> +
> + if (!file_exists(git_path("commondir")))
> + return;
> + strbuf_addf(&sb, "%s/", get_git_dir());
> + len = sb.len;
> + for (p = list; *p; p++) {
> + strbuf_setlen(&sb, len);
> + strbuf_addstr(&sb, *p);
> + if (file_exists(sb.buf))
> + report_garbage("unused in linked checkout", sb.buf);
> + }
> + strbuf_release(&sb);
> +}
> +
> static char const * const count_objects_usage[] = {
> N_("git count-objects [-v] [-H | --human-readable]"),
> NULL
> @@ -102,8 +135,10 @@ int cmd_count_objects(int argc, const char **argv, const char *prefix)
> /* we do not take arguments other than flags for now */
> if (argc)
> usage_with_options(count_objects_usage, opts);
> - if (verbose)
> + if (verbose) {
> report_garbage = real_report_garbage;
> + report_linked_checkout_garbage();
> + }
> memcpy(path, objdir, len);
> if (len && objdir[len-1] != '/')
> path[len++] = '/';
> diff --git a/path.c b/path.c
> index 47383ff..2e6035d 100644
> --- a/path.c
> +++ b/path.c
> @@ -92,6 +92,10 @@ static void replace_dir(struct strbuf *buf, int len, const char *newdir)
>
> static void update_common_dir(struct strbuf *buf, int git_dir_len)
> {
> + /*
> + * Remember to report_linked_checkout_garbage()
> + * builtin/count-objects.c
> + */I couldn't figure out why this comment was telling me to remember to report "linked checkout garbage" until I realized that you omitted the word "update" (as in "remember to update"). It might be clearer to say something along these lines:
Keep synchronized with related list in
builtin/count-objects.c:report_linked_checkout_garbage().Is it not possible or just too much of a hassle to maintain this list in just one place, as in a header which is included by these two files? The exceptions, such as 'log' and 'gc.pid', could be marked by a special character in the entry ("!gc.pid", for example) or any such scheme.
Show 5 quoted lines
> const char *common_dir_list[] = {
> "branches", "hooks", "info", "logs", "lost-found", "modules",
> "objects", "refs", "remotes", "repos", "rr-cache", "svn",
> --
> 1.9.0.40.gaa8c3ea