[xine-devel] this_gen vs directly this

[Diego 'F. <fla...@ge...>]

So, tonight I suffered from a few debugging problems while trying to implem=
ent=20
the mmapped file support (it's comfortable when you try 6 files, it works,=
=20
you try the 7th, it fails, eh? :) ). Luckily now it's all fixed, but during=
=20
the debug session, I've started wondering something.

Most of the plugin functions in xine are declared this way:

static off_t file_plugin_read (input_plugin_t *this_gen, char *buf, off_t l=
en)=20
{
  file_input_plugin_t *this =3D (file_input_plugin_t *) this_gen;

Unfortunately this is difficult to debug in gdb, because the minimum=20
optimisation you add, this will be removed out of debugger's view.

I suppose this declaration is to force the first parameter to conform so th=
at=20
when using the function's pointer it doesn't collide with the declaration o=
f=20
the function. This is all well and good, but I'm not sure if it's worth the=
=20
hassle of the lost debugging.

Declaring the one above as

static off_t file_plugin_read (file_input_plugin_t *this, char *buf, off_t=
=20
len) {

makes sure that the symbol "this" is almost always visible on the debugger,=
=20
and only spawns an extra warning that can be easily made silent, by casting=
=20
the function when passing the function's parameter to the struct.

Is there anything else for which the first notation is needed in xine?

=2D-=20
Diego "Flameeyes" Petten=F2 - http://farragut.flameeyes.is-a-geek.org/
Gentoo/Alt lead, Gentoo/FreeBSD, Video, AMD64, Sound, PAM, KDE