[OpenAFS-port-darwin] Re: Example of the "correct" way to get tokens for Finder on login...

[Edward Moy]

I should make it clear that what I say and do is not Apple talking.   
I am just another interested developer, who happens to work at  
Apple.  I have been a long time proponent of AFS and I'm glad to see  
that Apple is trying to better support all our customers who use  
OpenAFS.

As for the particular change to fix the LWP stack misalignment  
problem, the issue is to keep the stack pointer when coming into a  
routine equal a modulo 16 of 12.  So with a pointer aligned to 16,  
whether we add 12 or subtract 4 doesn't make that much difference.  I  
take the extra effort to insure 16-byte alignment (like changing the  
default stack size from 1000 to 1008) and to align the pointer from  
malloc (which actually isn't necessary, since malloc in Mac OS X  
always returns a pointer that is 16-byte aligned).

The rest of my changes are for building 2-way universal, fixing an  
fstab problem in the pthread version of fileserver, improving the  
packaging mechanism and assorted bug and compiler warning fixes.

Edward Moy
Apple

On Mar 9, 2006, at 1:53 PM, Everette Allen wrote:

> I agree but with the difference that I would like the OpenAFS  
> community and GateKeepers to agree with Apple on the choice of  
> mechanism.  For Apple to say "Thou shall..." do one thing when  
> another solution might fit better in the overall plan and upkeep of  
> OpenAFS is really not in any of our best interests.  Both sides  
> need to be able to sustain and support this important feature long  
> term.
>
> Christopher D. Clausen wrote:
>> Everette Allen <Everette_Allen@ncsu.edu> wrote:
>>> So back to the real question... the window folks have a "blessed" by
>>> the openafs community and MS mechanism to acquire tokens usable to
>>> the gui and the MacOS platform does not, what do we need/want and  
>>> how do we go
>>> about getting to this point?   Right now we have a mishmash of  
>>> cobbled
>>> together mechanisms which may or may not survive even minor OS  
>>> updates
>>> and that needs to change...
>> I would suggest talking with Apple to make sure that there is  
>> "blessed" way of doing what we need done before anyone attempts to  
>> actually write it.  We have "a mishmash of cobbled together  
>> mechanisms which may or may not survive even minor OS updates"  
>> because there does not appear to be a method for doing what the  
>> aklogin plugin needs to do through Apple approved methods.
>> <<CDC
>
> -- 
> Everette Gray Allen		Systems Programmer II
> ITD Computing Services	Macintosh Support Specialist
> 2620 Hillsborough St, Campus Box 7109
> Raleigh, NC 27695-7109  AIM: EveretteAlln
> 919-515-4558		Everette_Allen@ncsu.edu
> _______________________________________________
> port-darwin mailing list
> port-darwin@openafs.org
> https://lists.openafs.org/mailman/listinfo/port-darwin