git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Erik Faye-Lund <kusmabite@gmail.com>
To: Michael Weiser <M.Weiser@science-computing.de>
Cc: git@vger.kernel.org
Subject: Re: [PATCH] Extend runtime prefix computation
Date: Fri, 30 Nov 2012 11:20:52 +0100	[thread overview]
Message-ID: <CABPQNSYhscHdnTFLye=oif_R84kpdaVsrCK+-174v7Ugrae_yQ@mail.gmail.com> (raw)
In-Reply-To: <20121127163004.GC7499@science-computing.de>

On Tue, Nov 27, 2012 at 5:30 PM, Michael Weiser
<M.Weiser@science-computing.de> wrote:
> Support determining the binaries' installation path at runtime even if
> called without any path components (i.e. via search path). Implement
> fallback to compiled-in prefix if determination fails or is impossible.
>
> Signed-off-by: Michael Weiser <weiser@science-computing.de>
> ---
> - Has two very minor memory leaks - function is called only once per
>   program execution. Do we care? Alternative: Use static buffer instead.
>
>  exec_cmd.c |   68 ++++++++++++++++++++++++++++++++++++++++++++++-------------
>  1 files changed, 53 insertions(+), 15 deletions(-)
>
> diff --git a/exec_cmd.c b/exec_cmd.c
> index 125fa6f..d50d7f8 100644
> --- a/exec_cmd.c
> +++ b/exec_cmd.c
> @@ -4,28 +4,22 @@
>  #define MAX_ARGS       32
>
>  static const char *argv_exec_path;
> -static const char *argv0_path;
> +static const char *argv0_path = NULL;
>
>  const char *system_path(const char *path)
>  {
> -#ifdef RUNTIME_PREFIX
> -       static const char *prefix;
> -#else
>         static const char *prefix = PREFIX;
> -#endif
>         struct strbuf d = STRBUF_INIT;
>
>         if (is_absolute_path(path))
>                 return path;
>
>  #ifdef RUNTIME_PREFIX
> -       assert(argv0_path);
> -       assert(is_absolute_path(argv0_path));
> -
> -       if (!prefix &&
> -           !(prefix = strip_path_suffix(argv0_path, GIT_EXEC_PATH)) &&
> -           !(prefix = strip_path_suffix(argv0_path, BINDIR)) &&
> -           !(prefix = strip_path_suffix(argv0_path, "git"))) {
> +       if (!argv0_path ||
> +           !is_absolute_path(argv0_path) ||
> +           (!(prefix = strip_path_suffix(argv0_path, GIT_EXEC_PATH)) &&
> +            !(prefix = strip_path_suffix(argv0_path, BINDIR)) &&
> +            !(prefix = strip_path_suffix(argv0_path, "git")))) {
>                 prefix = PREFIX;
>                 trace_printf("RUNTIME_PREFIX requested, "
>                                 "but prefix computation failed.  "
> @@ -41,20 +35,64 @@ const char *system_path(const char *path)
>  const char *git_extract_argv0_path(const char *argv0)
>  {
>         const char *slash;
> +       char *abs_argv0 = NULL;
>
>         if (!argv0 || !*argv0)
>                 return NULL;
>         slash = argv0 + strlen(argv0);
>
> +       /* walk to the first slash from the end */
>         while (argv0 <= slash && !is_dir_sep(*slash))
>                 slash--;
>
> +       /* if there was a slash ... */
>         if (slash >= argv0) {
> -               argv0_path = xstrndup(argv0, slash - argv0);
> -               return slash + 1;
> +               /* ... it's either an absolute path */
> +               if (is_absolute_path(argv0)) {
> +                       /* FIXME: memory leak here */
> +                       argv0_path = xstrndup(argv0, slash - argv0);
> +                       return slash + 1;
> +               }
> +
> +               /* ... or a relative path, in which case we have to make it
> +                * absolute first and do the whole thing again */
> +               abs_argv0 = xstrdup(real_path(argv0));
> +       } else {
> +               /* argv0 is no path at all, just a name. Resolve it into a
> +                * path. Unfortunately, this gets system specific. */
> +#if defined(__linux__)
> +               struct stat st;
> +               if (!stat("/proc/self/exe", &st)) {
> +                       abs_argv0 = xstrdup(real_path("/proc/self/exe"));
> +               }
> +#elif defined(__APPLE__)
> +               /* Mac OS X has realpath, which incidentally allocates its own
> +                * memory, which in turn is why we do all the xstrdup's in the
> +                * other cases. */
> +               abs_argv0 = realpath(argv0, NULL);
> +#endif

...perhaps a "GetModuleFileName(NULL, ...)" for Windows is in place here?

  reply	other threads:[~2012-11-30 10:29 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-11-27 16:30 [PATCH] Extend runtime prefix computation Michael Weiser
2012-11-30 10:20 ` Erik Faye-Lund [this message]
2012-11-30 10:45   ` Michael Weiser
2013-03-05 11:58 ` Michael Weiser
2013-03-05 16:13   ` Junio C Hamano
2013-03-06  8:19     ` Michael Weiser
2013-04-16 14:56       ` Michael Weiser
2013-04-16 18:23         ` Junio C Hamano
2013-04-16 15:18       ` Erik Faye-Lund
2013-04-17  6:06         ` Michael Weiser
2013-10-04 13:32           ` Michael Weiser
  -- strict thread matches above, loose matches on Subject: below --
2016-04-15 14:30 Michael Weiser
2016-04-15 16:43 ` Junio C Hamano
2016-04-18  7:20   ` Johannes Schindelin
2016-04-20 17:52   ` Michael Weiser

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='CABPQNSYhscHdnTFLye=oif_R84kpdaVsrCK+-174v7Ugrae_yQ@mail.gmail.com' \
    --to=kusmabite@gmail.com \
    --cc=M.Weiser@science-computing.de \
    --cc=git@vger.kernel.org \
    /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;
as well as URLs for NNTP newsgroup(s).