Re: [PATCH] gc: remove gc.pid file at end of execution

[Junio C Hamano]

Jonathan Nieder  writes:

> Matthieu Moy wrote:
>
>> This file isn't really harmful, but isn't useful either, and can create
>> minor annoyance for the user:
>
> Would something like the following make sense, to ensure the gc.pid file is
> always removed on normal exit?

Has anything further happened to this discussion?
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] gc: remove gc.pid file at end of execution

[Duy Nguyen]

On Sat, Sep 28, 2013 at 7:33 AM, Jonathan Nieder  wrote:
> Matthieu Moy wrote:
>
>> This file isn't really harmful, but isn't useful either, and can create
>> minor annoyance for the user:
>
> Would something like the following make sense, to ensure the gc.pid file is
> always removed on normal exit?
>
> Signed-off-by: Jonathan Nieder 
>
> diff --git c/builtin/gc.c i/builtin/gc.c
> index 6e0d81a..fbbc144 100644
> --- c/builtin/gc.c
> +++ i/builtin/gc.c
> @@ -14,6 +14,7 @@
>  #include "cache.h"
>  #include "parse-options.h"
>  #include "run-command.h"
> +#include "sigchain.h"
>  #include "argv-array.h"
>
>  #define FAILED_RUN "failed to run %s"
> @@ -167,6 +168,21 @@ static int need_to_gc(void)
> return 1;
>  }
>
> +static char *pidfile;
> +
> +static void remove_pidfile(void)
> +{
> +   if (pidfile)
> +   unlink(pidfile);
> +}
> +
> +static void remove_pidfile_on_signal(int signo)
> +{
> +   remove_pidfile();
> +   sigchain_pop(signo);
> +   raise(signo);
> +}
> +
>  /* return NULL on success, else hostname running the gc */
>  static const char *lock_repo_for_gc(int force, pid_t* ret_pid)
>  {
> @@ -179,13 +195,19 @@ static const char *lock_repo_for_gc(int force, pid_t* 
> ret_pid)
> FILE *fp;
> int fd, should_exit;
>
> +   if (pidfile)
> +   /* already locked */
> +   return NULL;
> +
> if (gethostname(my_host, sizeof(my_host)))
> strcpy(my_host, "unknown");
>
> -   fd = hold_lock_file_for_update(&lock, git_path("gc.pid"),
> +   pidfile = git_pathdup("gc.pid");
> +
> +   fd = hold_lock_file_for_update(&lock, pidfile,
>LOCK_DIE_ON_ERROR);
> if (!force) {
> -   fp = fopen(git_path("gc.pid"), "r");
> +   fp = fopen(pidfile, "r");
> memset(locking_host, 0, sizeof(locking_host));
> should_exit =
> fp != NULL &&
> @@ -208,6 +230,7 @@ static const char *lock_repo_for_gc(int force, pid_t* 
> ret_pid)
> if (should_exit) {
> if (fd >= 0)
> rollback_lock_file(&lock);
> +   pidfile = NULL;
> *ret_pid = pid;
> return locking_host;
> }
> @@ -219,6 +242,9 @@ static const char *lock_repo_for_gc(int force, pid_t* 
> ret_pid)
> strbuf_release(&sb);

It may be a bit simpler to delay setting pidfile until we get here.
lock.filename still contains gc.pid until commit_lock_file is called.

> commit_lock_file(&lock);
>
> +   sigchain_push_common(remove_pidfile_on_signal);
> +   atexit(remove_pidfile);
> +
> return NULL;
>  }
-- 
Duy
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] gc: remove gc.pid file at end of execution

[Jonathan Nieder]

Matthieu Moy wrote:

> This file isn't really harmful, but isn't useful either, and can create
> minor annoyance for the user:

Would something like the following make sense, to ensure the gc.pid file is
always removed on normal exit?

Signed-off-by: Jonathan Nieder 

diff --git c/builtin/gc.c i/builtin/gc.c
index 6e0d81a..fbbc144 100644
--- c/builtin/gc.c
+++ i/builtin/gc.c
@@ -14,6 +14,7 @@
 #include "cache.h"
 #include "parse-options.h"
 #include "run-command.h"
+#include "sigchain.h"
 #include "argv-array.h"
 
 #define FAILED_RUN "failed to run %s"
@@ -167,6 +168,21 @@ static int need_to_gc(void)
return 1;
 }
 
+static char *pidfile;
+
+static void remove_pidfile(void)
+{
+   if (pidfile)
+   unlink(pidfile);
+}
+
+static void remove_pidfile_on_signal(int signo)
+{
+   remove_pidfile();
+   sigchain_pop(signo);
+   raise(signo);
+}
+
 /* return NULL on success, else hostname running the gc */
 static const char *lock_repo_for_gc(int force, pid_t* ret_pid)
 {
@@ -179,13 +195,19 @@ static const char *lock_repo_for_gc(int force, pid_t* 
ret_pid)
FILE *fp;
int fd, should_exit;
 
+   if (pidfile)
+   /* already locked */
+   return NULL;
+
if (gethostname(my_host, sizeof(my_host)))
strcpy(my_host, "unknown");
 
-   fd = hold_lock_file_for_update(&lock, git_path("gc.pid"),
+   pidfile = git_pathdup("gc.pid");
+
+   fd = hold_lock_file_for_update(&lock, pidfile,
   LOCK_DIE_ON_ERROR);
if (!force) {
-   fp = fopen(git_path("gc.pid"), "r");
+   fp = fopen(pidfile, "r");
memset(locking_host, 0, sizeof(locking_host));
should_exit =
fp != NULL &&
@@ -208,6 +230,7 @@ static const char *lock_repo_for_gc(int force, pid_t* 
ret_pid)
if (should_exit) {
if (fd >= 0)
rollback_lock_file(&lock);
+   pidfile = NULL;
*ret_pid = pid;
return locking_host;
}
@@ -219,6 +242,9 @@ static const char *lock_repo_for_gc(int force, pid_t* 
ret_pid)
strbuf_release(&sb);
commit_lock_file(&lock);
 
+   sigchain_push_common(remove_pidfile_on_signal);
+   atexit(remove_pidfile);
+
return NULL;
 }
 
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html