[syslinux] Allowed code pages and encodings to write f0.txt through f1.txt?

[Thomas Schmitt]

Hi,

> > The existence of Joliet in an ISO image is announced by [...]
> > One would have to decide early, which one to use, if both are
> > present.

> Quite correct.

"Quite" ? Did i misrepresent some facts ?


> We would either go the Linux route of Rock Ridge > Joliet > plain 9660

One would have to decide which of the two root records to use.

It is easy to let Rock Ridge override plain 9660 on file-by-file
decision.
But to determine the Joliet companion of a plain 9660 name would need
a tree lookup in the Joliet tree. The only reliable key that links
both trees would be the block addresses of data files.
All other file types would not unambigously find their counterparts
in the Joliet tree. This includes directories and could make lookup
quite tricky.


> or be lazy

Rock-Ridge-only looks comfortable. Rock Ridge is well suitable for
UTF-8, because it makes no assumptions at all about the character
set or the permissible byte values. Everything except NUL and '/'
should be ok. (NUL is cumbersome for C programs. I wonder what weird
effects '/' could have on unaware readers.)


I am pondering whether i should contribute a Rock Ridge name
retriever function. But building a test frame for iso9660.c would
be more effort than changing its code. Learning how to build ISOLINUX
and to create a test case might become an even longer endeavour ...

Is anybody willing to do the testing part of Rock Ridge in ISOLINUX ?
(With filenames which would not work with current ISOLINUX.)
((And then submit the patch in our both name ?))


Have a nice day :)

Thomas