Is there any advantage of such coding-style?

Question

|There is this function in gestremer matroska demux plugin:

gboolean
gst_matroska_demux_plugin_init (GstPlugin * plugin)
{
  /* parser helper separate debug */
  GST_DEBUG_CATEGORY_INIT (ebmlread_debug, "ebmlread",
      0, "EBML stream helper class");

  /* create an elementfactory for the matroska_demux element */
  if (!gst_element_register (plugin, "matroskademux",
          GST_RANK_PRIMARY, GST_TYPE_MATROSKA_DEMUX))
    return FALSE;

  return TRUE;
}

Now gst_element_register() is type of

gboolean            gst_element_register                (GstPlugin *plugin,
                                                         const gchar *name,
                                                         guint rank,
                                                         GType type);
Returns :
    TRUE, if the registering succeeded, FALSE on error

Then why not write it in the following way?

 gboolean
    gst_matroska_demux_plugin_init (GstPlugin * plugin)
    {
      /* parser helper separate debug */
      GST_DEBUG_CATEGORY_INIT (ebmlread_debug, "ebmlread",
          0, "EBML stream helper class");

      /* create an elementfactory for the matroska_demux element */
      return gst_element_register (plugin, "matroskademux",
              GST_RANK_PRIMARY, GST_TYPE_MATROSKA_DEMUX))      
    }

Answer 1

It's part of a pattern.

if (!some_function(...))
    return false;
if (!other_function(...))
    return false;

return true;

Whoever wrote it decided not to change the pattern just because there's only one function call. Ultimately it's a matter of taste.

Answer 2

As such there is no problem with the code. At least I wont penalize if anyone uses any of the mentioned snippets.

These are the reasons that I think are the cause:

• The foremost is readability. Now looking at the section code, I know what is success and what is not. As there could be apis, which behaves inversely (even in C).
• Second reason happens, when during testing you want to print some log statements. Now if you return directly you cannot log the return value or any debug info. You can say after code is freed, one should remove it, but then for next release you might have to put debug statements again. So it stays there.
• Third reason could be that some code is removed during development.

Conclusion is what I said in the beginning. It does not matter as long as code is easy to understand. Regarding some optimization gains, I think for this compilers are intelligent enough to take care.