Re: [Bug-xorriso] bug with handling symbolic links?

[Thomas Schmitt]

Hi,

> thanks for the ultrafast diagnosis and reply.

Regrettably it is not so easy to change libisofs
to comply with Linux isofs expectations. The original
libisofs programmer, Vreixo Formoso Lopes, took some
advantage from his (probably correct) interpretation
of SUSP specs. Now it gets more complicated.

I am currently trying to find out whether mkisofs
exercised the kernel with the cheapest possible
workaround i could invent yet. Probably i will have
to implement a more costly variant.

It might last a few days until i can show something
to test.


> > So i will have to change libisofs to fulfill
> > the (probably incorrect) expectations of Linux.
> ... and probably also those of iso{info,vfy} from
> cdr{tools,kit} at the same time?

Yeah. They form a club. I wonder whether their author
knew more than Vreixo and me. Linux fs/isofs, mkisofs,
and isoinfo all were founded by Eric Youngdale in the
early 1990s.
His specs seem to have mentioned entry RR, of which
no trace is to see in the SUSP and RRIP specs which
i could find.
Maybe there was also the prescription to keep the
SUSP continuation data of a file in a single block
and to address this block entirely by the Block Location
number field of CE and not in part by the byte Offset
field of CE.


Have a nice day :)

Thomas