Re: link(2) allows directory hard links, man page says otherwise
Re: link(2) allows directory hard links, man page says otherwise
- Subject: Re: link(2) allows directory hard links, man page says otherwise
- From: Rosyna <email@hidden>
- Date: Wed, 12 Dec 2007 17:41:28 -0700
This is true. And directory hard links are a concept I do not like for
various reasons. So, I was really pleased when people at WWDC and
other places repeatedly said (or strongly implied) Time Machine used a
private function to make directory hard links, that this SPI would not
be exposed, and developers should not be making directory hard links.
But alas, if some developer has code that depends on link() failing
with EPERM for a directory (they don't check the path first), their
code will magically start working on HFS+ on 10.5.
My hope was that time machine would call some super private function
(like hfs_super_private_link) that exposes hfs_vnop_link() to do its
dirty deeds that no one else could, should, or can call.
It's not my intention to ever intentionally create directory hard links.
On Dec 12, 2007, at 4:37 PM, Jordan K. Hubbard wrote:
You're not supposed to notice that directory hard links are now
allowed (for Time Machine's exclusive use) so it's neither an error
in the documentation or an error in the function. The error is that
you're being too nosy. :-)
- Jordan
On Dec 12, 2007, at 3:12 PM, Rosyna wrote:
The actual implementation of link() on 10.5 allows for directory
hard links on HFS+ as long as they follow these rules:
---
Sincerely,
Rosyna Keller
Technical Support/Carbon troll/Always needs a hug
Unsanity: Unsane Tools for Insanely Great People
It's either this, or imagining Phil Schiller in a thong.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden