Prev: Non-Flash flashing ads
Next: Low end HP printers & OS X 10.6
From: R on 28 Apr 2010 06:45
Chris Ridd <chrisridd(a)mac.com> wrote:
> Yup, when I tried to reproduce this the perms on /Users/shared stayed
> the same. But then I could also rm the root_file.
>
> dtracing the kernel's ACL(etc) code might be the only way to discover
> what's happening.

Do you think that is something within our reach or would it be
best left to Apple?

Btw and in case you wonder:

bash: uname -v
Darwin Kernel Version 10.3.0: Fri Feb 26 11:57:13 PST 2010;
root:xnu-1504.3.12~1/RELEASE_X86_64

I've reported this as a potential bug to Apple.
From: Chris Ridd on 28 Apr 2010 06:55
On 2010-04-28 11:45:47 +0100, R said:

> Chris Ridd <chrisridd(a)mac.com> wrote:
>> Yup, when I tried to reproduce this the perms on /Users/shared stayed
>> the same. But then I could also rm the root_file.
>>
>> dtracing the kernel's ACL(etc) code might be the only way to discover
>> what's happening.
>
> Do you think that is something within our reach or would it be
> best left to Apple?

The kernel code's open source, and there are dtrace providers in the kernel.

>
> Btw and in case you wonder:
>
> bash: uname -v
> Darwin Kernel Version 10.3.0: Fri Feb 26 11:57:13 PST 2010;
> root:xnu-1504.3.12~1/RELEASE_X86_64

You're booting a 64-bit kernel?

>
> I've reported this as a potential bug to Apple.


--
Chris

From: R on 28 Apr 2010 07:02
Jaimie Vandenbergh <jaimie(a)sometimes.sessile.org> wrote:

> The "escalation" survives a closure of Terminal.app and restart. I
> haven't tried a logout/login.

I tried that a couple of times. Sometimes it survives the logout
and login, sometimes not. I'm not sure if the logging out is
involved though. There could be other unrelated causes, e.g.,
something timing out.

> This is a new iMac with the 10.6.2 upgraded to 10.6.3, and an old
> (originally tiger) user account which has been Migrated across half a
> dozen machines...

Installed to an empty disk here (i.e., fresh install).

Other observations: the problem occurs in '/tmp' as well as in
'/Users/Shared'. I have observed that the problem cannot be
replicated in either '/tmp' or '/Users/Shared' on another HFS+
formatted partition I mounted under /Volumes.

Next step might be to create a new sticky bit directory to
experiment on.
From: R on 28 Apr 2010 07:03
Chris Ridd <chrisridd(a)mac.com> wrote:

> > bash: uname -v
> > Darwin Kernel Version 10.3.0: Fri Feb 26 11:57:13 PST 2010;
> > root:xnu-1504.3.12~1/RELEASE_X86_64
>
> You're booting a 64-bit kernel?

Yes, sir!
From: Jaimie Vandenbergh on 28 Apr 2010 08:29
On Wed, 28 Apr 2010 12:03:24 +0100, me32(a)privacy.net (R) wrote:

>Chris Ridd <chrisridd(a)mac.com> wrote:
>
>> > bash: uname -v
>> > Darwin Kernel Version 10.3.0: Fri Feb 26 11:57:13 PST 2010;
>> > root:xnu-1504.3.12~1/RELEASE_X86_64
>>
>> You're booting a 64-bit kernel?
>
>Yes, sir!

I'm not, so that doesn't have bearing on the stickies.

Cheers - Jaime
--
"Every Little Thing She Does Is Sufficiently Advanced Technology"
First  |  Prev  |  Next  |  Last
Pages: 1 2 3 4 5
Prev: Non-Flash flashing ads
Next: Low end HP printers & OS X 10.6