On Thu, Apr 19, 2001 at 03:31:08PM -0800, Ethan Benson wrote: > On Thu, Apr 19, 2001 at 10:51:17AM -0500, W. Crowshaw wrote: > > Actually, it doesn't hurt anything if the MacOS HFS FS is mounted on the > > kernel > > and hmount'ed with hfsutils. I access the HFS FS both mounted and hmounted > > all the time. > > then you have been lucky. the kernel will NOT be aware of the changes > to the filesystem that your making behind its back with hfsutils, > since hfsutils manipulates the raw device (/dev/[sh]d?[0-9]*). > modifying a read-write mounted filesystem with hfsutils is just asking > for serious filesystem corruption and even possibly a kernel > malfunction. >
I'm not reckless enough to actually change anything on a hmounted HFS partition when it is also mounted on the kernel. If the partition is mounted read-only, I have found no harm in hmounting it as well, viewing it with hls, and hcopy'ing files from it to an ext2 FS. As a matter of fact, I do this everyday as a way to backup important WORD documents on the Mac side. To be successful (not necessarily with WORD files which are data fork only), you just have to macbinarize them with hcopy -m before they are copied to an ext2 partition. With cron, a properly written shell script, and hfsutils, it works as a backup solution. Writing them back to an hmounted HFS volume is a another matter completely. But it can be done on an empty (for safety's sake) HFS hmounted zip floppy. These newly restored files can, will, and do show up under the MacOS on next boot. I have done this on a regular basis without losing files, f'ing up partitions, crashing the kernel, killing the MacOS, burning up my hardware, or ending life as we know it. fyi, -- wcrowshaw
