Re: [Libcdio-devel] CD_MSF_FORMAT vs LBA on NetBSD

[Thomas Schmitt]

Hi,

> To clarify, before the fix I posted in my previous email:

So was the shown cd-info result from the intermediate state before
  
https://github.com/vext01/libcdio/commit/539c52488506c5bacdfebbcafc5406034a0738fd
?
(Pity, i had it in mind as reason for the cdparanoia complaint.)


But why then didn't this yield an error message ?
  TOTAL_TRACKS + 1 = 6, first_track = 7 , 
  so track_num < first_track
which should have caused get_track_msf_netbsd() to return
CDIO_INVALID_TRACK.

So this error indication gets lost on its way to cd-info.c ?

Well, its a bit more off the road. In src/cd-info.c i read:

    if (!cdio_get_track_msf(p_cdio, i, &msf)) {
      err_exit("cdio_track_msf for track %i failed, I give up.\n", i);
    }

lib/driver/track.c has
  if (p_cdio->op.get_track_msf) {
    return p_cdio->op.get_track_msf (p_cdio->env, u_track, msf);

and netbsd.c has

  static cdio_funcs_t _funcs = {
     ...
     .get_track_msf = get_track_msf_netbsd,

which returns in the case of bad track number neither "true" nor "false",
but rather CDIO_INVALID_TRACK.

include/cdio/track.h :
   extern enum cdio_track_enums {
      ...
      CDIO_INVALID_TRACK       = 0xFF, /* ... */

include/cdio/types.h
  #define true 1
  #define false 0

=======================================================================

So get_track_msf_netbsd() must not return CDIO_INVALID_TRACK but rather
false in case of error.

Alternatively cdio_get_track_msf() could check for values other than
true and false. But CDIO_INVALID_TRACK seems to be intended as pseudo-
track number, not as success indicator. Further get_track_msf_netbsd()
is indeed declared as "bool".


Have a nice day :)

Thomas