Re: cp'ing passive translators
The question I have is general and I think it's related to this
thread.
In my opinion, the passive translators used for auto-mounting
filesystems are very different from the ones used for e.g. device
files. Sure, technically they are the same, but the policy when to
use the translator and when to use the underlying file with it's
passive translator setting differs, I think.
Imho, the complete filesystem tree consists of the root filesystem
plus all filesystems installed as passive translators. They are
simply started when they are accessed, but this is completely
transparent to the user. The result of not automatically following
them (for example during a recursive copy) would be that this
filesystem tree is not traversed completely.
On the other hand, there are those translator, which don't provide a
filesystem and starting them or not wouldn't extend or shrink the
filesystem tree.
If i would like to clone a GNU system via archiving the filesystem
tree and extracting it later somewhere else, I would expect to also
have /home archived although it's a different partition than the root
filesystem, it's a passive translator sitting on /home. But I would
also expect that /dev/zero would be simply a passive translated file
on the destination system.
It would be great, if somebody could enlighten me here. What am I
missing? What were the design decisions in this respect?
Thanks,
moritz
--
moritz@duesseldorf.ccc.de - http://duesseldorf.ccc.de/~moritz/
GPG fingerprint = 3A14 3923 15BE FD57 FC06 B501 0841 2D7B 6F98 4199
Reply to: