Re: [PATCH 1/1] change contract between system_path and it's callers

[Junio C Hamano <gitster@pobox.com>]

Alexander Kuleshov <kuleshovmail@gmail.com> writes:

> Now system_path returns path which is allocated string to callers;
> It prevents memory leaks in some places. All callers of system_path
> are owners of path string and they must release it.
>
> Signed-off-by: Alexander Kuleshov <kuleshovmail@gmail.com>
> ---

Comparing this with what I sent out...

>  builtin/help.c | 10 +++++++---
>  exec_cmd.c     | 17 +++++++++--------
>  exec_cmd.h     |  4 ++--
>  git.c          | 16 ++++++++++++----
>  4 files changed, 30 insertions(+), 17 deletions(-)
>
> @@ -372,7 +373,9 @@ static void show_man_page(const char *git_cmd)
>  static void show_info_page(const char *git_cmd)
>  {
>  	const char *page = cmd_to_page(git_cmd);
> -	setenv("INFOPATH", system_path(GIT_INFO_PATH), 1);
> +	char *git_info_path = system_path(GIT_INFO_PATH);
> +	setenv("INFOPATH", git_info_path, 1);
> +	free(git_info_path);

We are just about to exec; does this warrant the code churn?

>  	execlp("info", "info", "gitman", page, (char *)NULL);
>  	die(_("no info viewer handled the request"));

> @@ -34,8 +34,7 @@ const char *system_path(const char *path)
>  #endif
>  
>  	strbuf_addf(&d, "%s/%s", prefix, path);
> -	path = strbuf_detach(&d, NULL);
> -	return path;
> +	return d.buf;

These happens to be the same with the current strbuf implementation,
but it is a good manner to use strbuf_detach(&d, NULL) here.  We
don't know what other de-initialization tomorrow's implementation of
the strbuf API may have to do in strbuf_detach().

> @@ -68,16 +67,16 @@ void git_set_argv_exec_path(const char *exec_path)
>  
>  
>  /* Returns the highest-priority, location to look for git programs. */
> -const char *git_exec_path(void)
> +char *git_exec_path(void)
>  {
>  	const char *env;
>  
>  	if (argv_exec_path)
> -		return argv_exec_path;
> +		return strdup(argv_exec_path);
>  
>  	env = getenv(EXEC_PATH_ENVIRONMENT);
>  	if (env && *env) {
> -		return env;
> +		return strdup(env);
>  	}

Now you are making callers of git_exec_path() responsible for
freeing the result they receive.

git_exec_path() may be called quite a lot, which means we may end up
calling system_path() many times during the life of a process
without freeing its return value, so this change may be worth doing,
but this patch is insufficient, isn't it?

You just added load_command_list() in help.c a new leak or two, for
example.  There probably are other callers of this function but I
don't have time to look at all of them myself right now.

> @@ -95,8 +94,10 @@ void setup_path(void)
>  {
>  	const char *old_path = getenv("PATH");
>  	struct strbuf new_path = STRBUF_INIT;
> +	char* exec_path = git_exec_path();
>  
> -	add_path(&new_path, git_exec_path());
> +	add_path(&new_path, exec_path);
> +	free(exec_path);
>  	add_path(&new_path, argv0_path);

This part by itself is good, provided if we make it the caller's
responsiblity to free string returned by git_exec_path().

> diff --git a/git.c b/git.c
> index 82d7a1c..d01c4f1 100644
> --- a/git.c
> +++ b/git.c
> @@ -95,17 +95,25 @@ static int handle_options(const char ***argv, int *argc, int *envchanged)
>  			if (*cmd == '=')
>  				git_set_argv_exec_path(cmd + 1);
>  			else {
> -				puts(git_exec_path());
> +				char *exec_path = git_exec_path();
> +				puts(exec_path);
> +				free(exec_path);
>  				exit(0);
>  			}
>  		} else if (!strcmp(cmd, "--html-path")) {
> -			puts(system_path(GIT_HTML_PATH));
> +			char *git_html_path = system_path(GIT_HTML_PATH);
> +			puts(git_html_path);
> +			free(git_html_path);
>  			exit(0);
>  		} else if (!strcmp(cmd, "--man-path")) {
> -			puts(system_path(GIT_MAN_PATH));
> +			char *git_man_path = system_path(GIT_MAN_PATH);
> +			puts(git_man_path);
> +			free(git_man_path);
>  			exit(0);
>  		} else if (!strcmp(cmd, "--info-path")) {
> -			puts(system_path(GIT_INFO_PATH));
> +			char *git_info_path = system_path(GIT_INFO_PATH);
> +			puts(git_info_path);
> +			free(git_info_path);
>  			exit(0);
>  		} else if (!strcmp(cmd, "-p") || !strcmp(cmd, "--paginate")) {
>  			use_pager = 1;

None of these warrant the code churn, I would say.