On Oct 3, 2006, at 8:29 AM, Elad Efrat wrote:

> Jason Thorpe wrote:
>
>> Uh, I think this is a bit premature.  I don't think the discussion  
>> was
>> really over yet.
>>
>> I don't want to be stuck with p_flag and p_flag2 forever if we let  
>> this
>> sit.
>>
>> Please back it out.
>
> Back it out?

Yes, please back it out.  You've made an ABI change before I think it  
was ready, IMO.

> so we can sit and wait until someone designs and implements
> a proplib dictionary in struct proc, gets it approved, tested, and
> commited?

Uh, a proplib dictionary in struct proc is probably about 30 minutes  
worth of work (including compiling the kernel -- I would let you test  
it during the implementation of your feature since I don't have a  
direct use for the capability myself, at the moment).  Time well  
spent, I think.

> I'd rather not wait with features I want to implement until we have
> reached the ultimate way to store process flags. I'm certainly not  
> going
> to implement either a proplib dictionary or the other scheme  
> suggested;
> if you or someone else are, please say so, and I'll back it out.

I can certainly do this.  But I'd like to ask -- why are you saying  
that YOU would not implement the necessary (and simple) changes to add  
a properties dictionary to struct proc?

-- thorpej