[Debburn-devel] Minor problems with cdrkit-1.1.0

scdbackup at gmx.net scdbackup at gmx.net
Thu Dec 7 12:36:43 CET 2006


Hi,

shall i still do this :

> "make clean" and then run "make build/Makefile" and
> watch the log output. Does it find magic.h?

To my understanding there is a problem with an evolving
API of libmagic and its usage within genisoimage/apple.c
which is not backward compatible enough for my system.

After reading
  http://www.die.net/doc/linux/man/man3/libmagic.3.html
an advanced remedy seems to be this (in apple.c, line 65) :

  #ifdef USE_MAGIC
  #include <magic.h>
  magic_t magic_state = NULL;

  /* ts A61207 : Avoid complaints about missing symbol MAGIC_ERROR */
  #ifndef MAGIC_ERROR
  #define MAGIC_ERROR 0 /* not implemented in this version of libmagic */
  #endif

  #endif /* USE_MAGIC */

This compiles nicely on my SuSE 9.0 by "make clean ; make".
I lack of use cases for testing, though.

If open_magic() without MAGIC_ERROR is too daring one could
as well #undef USE_MAGIC in case of missing MAGIC_ERROR.


-------------------------------------------------------------------

About -stream-file-name and my own pathspec needs.
My dream pathspecs for genisoimage would cover:

Multiple input streams from non-disk sources (here estimated to
20 MB which shall get padded up)
    "=d:20M:/in/iso/file1=cat /tmp/pipe1"
    "=d:20M:/in/iso/file_from_stdin=cat"
The latter is equivalent to what i understand from man mkisofs:
    -stream-media-size 20M \
    -stream-file-name /in/iso/file_from_stdin

Encryption (with exactly predictable input size)
    "=d:321543:/in/iso/file_encrypted=encryption_program file_clear"

Cutting out pieces (exactly predicted size)
    "=d:2047G:/in/iso/piece_2_of_4=cut_program file_fat 2047G 4094G"

The leading "=d:" together with -graft-points could indicate
the new "dynamic" pathspecs. I believe this sub-namespace
is still unoccupied in traditional mkisofs.
Other "=..." would stay free for future definitions.


Have a nice day :)

Thomas




More information about the Debburn-devel mailing list