[EMAIL PROTECTED] wrote: >After the disk is written, I can mount the disk from Linux >(2.2.25, glibc 2.2.5) in iso9660 mode. However, for files that >originally have a Apple resource fork (from Netatalk), it seems >that the resource fork instead of the data fork is visible to >Linux. Stranger still, this does not happen if I mount the CD >image file instead of the CDROM disk. It seems that mkisofs >generates identical inode numbers for both forks (in iso9660 >mode), and the way the files are laid out are making at least >the Linux kernel confused.
I don't use mkisofs or Linux, but I do recall that the difference between a "resource fork" and a "data fork" in the ISO9660 format is merely the setting of a single bit. That bit indicates that the first extent of the file is a "resource fork". Otherwise it is a "data fork". So the question is whether mkisofs is not setting the bit or Linux is not skipping the first extent if the bit is set. An independent analysis tool would be required. I have one for VMS, but that will not help you if you have no VMS system. Try mounting the disc on a Macintosh. Try mounting the pure ISO9660 disc on a Macintosh. If the Macintosh gets both right (albeit without resource fork info in the second case) you problem is likely Linux. But that is what open source is about, right ? You can look for the checking of that bit and fix the code yourself. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

