On Wed, 26 Jan 2005, Dallman Ross wrote:
On Wed, Jan 26, 2005 at 11:19:09AM +0200, Udi Mottelo wrote:
On Tue, 25 Jan 2005, Dallman Ross wrote:
If the delivery is to a directory or directories -- either
maildir-, MH-, or traditional procmail dir-format -- then, yes.
If to flat folders (mbox-style files), then, no.
[...]
Thanks for the example, I didn't know it. But now after I
read the man again I think that you missed one more test:
[...]
The inode is same because there are no two copies
but hardlink. It can help for access by two ways
to the same message, for example:
vi $username/new/message or vi $familyname/new/message
but if you delete $username/new/message then
$familyname/new/message will not be exist!
It is hardlinked, yes. But your conclusion in the last sentence
just above is not correct. Deleting one copy of a hardlinked
file does leave the remaining cop(y|ies).
Sorry. you right. But still we can make damage by editing.
[...]
I haven't looked to see what procmail does on two different filesystems.
Let's try:
procmail: Rcfile: "two.rc"
procmail: Assigning "LASTFOLDER=foo/new/1106734272.20200_0.panix5.panix.com"
procmail: Linking to "/var/tmp/bar/new/1106734272.20200_1.panix5.panix.com"
procmail: Couldn't make link to
"/var/tmp/bar/new/1106734272.20200_1.panix5.panix.com"
Interesting; the second copy fails in that event.
Yes, because hardlink is two entries for one inode in one
table which doesn't exist in other filesystem. The softlink
is one entry which contains the address of the real file.
Bye,
Udi
--
dman
____________________________________________________________
procmail mailing list Procmail homepage: http://www.procmail.org/
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/procmail