Hi, Edd Barrett wrote: > I've finally found some time to look at this. The following diff seems > to work, although I don't think I have any CDs where the first track > number isn't 1 to test with.
Commercial test candidates would be second CDs of a multi-CD album. Self-made CDs can get arbitrary start track numbers by the .cue file format which is understood (among others) by cdrdao, cdrecord, wodim, cdrskin. (See "Example of a CDRWIN cue sheet file named NIGHTCATS.CUE" in https://dev.lovelyhq.com/libburnia/libburn/raw/master/doc/cdtext.txt where "TRACK 01 AUDIO" could be "TRACK 07 AUDIO" followed by tracks 08 and 09. Maximum permittable track number is 99.) cdrskin has an option cd_start_tno= for a run with just a few .wav files. Like: cdrskin -v dev=/dev/rcd0d cd_start_tno=7 trackA.wav trackB.wav trackC.wav > https://github.com/vext01/libcdio/commit/d04c48e36e7bb810dede415950a8cdf69e625e0a The changes look good so far. But a gesture in lib/driver/gnu_linux.c (and also lib/driver/solaris.c) still differs from the one in lib/driver/netbsd.c. ------------------- gnu_linux.c: if (i_track == CDIO_CDROM_LEADOUT_TRACK) i_track = p_env->gen.i_tracks + p_env->gen.i_first_track; ------------------- netbsd.c: #define TOTAL_TRACKS (_obj->tochdr.ending_track \ - _obj->tochdr.starting_track + 1) ... if (track_num == CDIO_CDROM_LEADOUT_TRACK) track_num = TOTAL_TRACKS + 1; TOTAL_TRACKS looks like the number of tracks on the CD. But both, gnu_linux.c and now also netbsd.c consider (_obj->gen.i_first_track + _obj->gen.i_tracks) as highest permissible value of the nominal track number (i_track and track_num, respectively). That's one higher than the highest nominal track number found on CD. So i guess that netbsd.c should now join gnu_linux.c in setting the nominal track number of the lead-out track to that value. Have a nice day :) Thomas
