[Libburn] Re: CVS version and test/burniso

[Derek Foreman]

Ooops, thanks.  Good catch.

mmc_write should actually be returning 0 for success.

(I'll do that in cvs)

On Fri, 10 Feb 2006, scdbackup at gmx.net wrote:

> Hi,
>
> test/burniso.c is not to blame.
>
> The CVS libburn of yesterday (after the version.h repair)
> does not burn, because BE_CANCELLED uses a value that is
> also issued by mmc_write() on success.
>
>
> I traced the problem to  libburn/sector.c:get_sector()
>
>                err = d->write(d, d->nwa, out);
>                if (err == BE_CANCELLED)
>                          return NULL;
>
> which already fails with SAO when called by  burn_write_track()
>                if (t->pregap2)
>                          for (i = 0; i < 150; i++)
>                                if (!sector_pregap(o, t->entry->point,
>                                                   t->entry->control, t->mode))
>                                          return 0;
>
> or in RAW mode when called by  burn_write_leadin()
>        while (count != 0) {
>                if (!sector_toc(o, s->track[0]->mode))
>                        return 0;
>
> in mmc_write() i read
>        if (cancelled)
>                return BE_CANCELLED;
>        ... successful operation ...
>        return 1;
>
> So the root of the problem is here
>   libburn/error.h:#define BE_CANCELLED 1
>
>
> We will need a different error number or a different
> ok indicator, as it seems.
> #define BE_CANCELLED 2 ... oh yeah. It works.
>
> Some day i will be a very good taxi driver within libburn :))
>
>
> A hint: libburn verbosity is still unusable with SAO
> because of a zillion messages already at level 1:
> "logical unit is in the process of becoming ready"
>
>
> Have a nice day :)
>
> Thomas
>
>
> _______________________________________________
> libburn mailing list
> libburn at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/libburn
>