[Libburn-hackers] Re: Issue burning CD-RW with unpredictable data size

Joris Dobbelsteen joris at familiedobbelsteen.nl
Sun Apr 4 06:00:16 CDT 2010

Thomas Schmitt wrote, On 26-03-10 09:25:
> Hi,
>> The test application works fine when burning to "stdio:/dev/null".
> Which one: yours , or my altered test/libburner ?

The example (altered test/libburner).

>> But when writing to CD-R or CD-RW it goes wrong (both in simulation and
>> normal writing). It continues until the end of the CD is reached. With
>> non-simulation information gets written and blanking (or chucking away) is
>> required.
>> - Lite-on DVDRW SHM-165P6S
> Neither good nor bad is known about this drive.
> But in general i would not expect the drive to
> be the culprit with surplus input.

I did get it working with cdrecord using the "AOpen" CD-writer drive a 
long time ago, so it is possible. However it seems libburn does not 
support it well.

> I rather suspect to find the decisive difference
> inside libburn. At about the middle of the layer
> stack, the stdio pseudo-drives split from MMC
> drives.
> For a closer examination we would need an output
> counter of the feeder and and input counter in
> libburn.
> We should also check what write type was chosen:
> SAO or TAO.
> This is reported to stderr if you set
>   burn_msgs_set_severities("NEVER", "DEBUG", "xyz");
> TAO reports:
>   xyz: DEBUG : TAO pre-track ...
> SAO :
>   xyz: DEBUG : SAO|RAW: Inquired nwa: ...

SAO|RAW is shown....
 From my notes, I believe you mentioned a few years back that SOA is 
required for burns of an unpredictable size.

> The feeder counter would be your job.
> Its value should be printed when the output
> of the feeder gets closed.
> I will now look for the place in libburn where
> to count the input and wher the decision about
> end-of-input is made.
>> It seems like a very good idea to debug the software on that system. If you
>> desire I can provide SSH access and a CD-RW in the box.
> Ok. Create a sandbox user and give him
> rw-permission to the drive.

Enclosed in a separate mail.

Thanks for the support!
And happy easter to all!


- Joris

More information about the Libburn-hackers mailing list