Re: non-caching MfD framework

2005-01-14 Thread Robert Story
On Tue, 11 Jan 2005 16:28:21 + Dave wrote: DS> RS> That's fine. But what I don't want to lose is the example code in DS> RS> container_load, which shows how to allocate the rowreq, set indexes DS> RS> and insert the rows. DS> DS> Well you wouldn't need to lose that - it's just that I'd expect

Re: non-caching MfD framework

2005-01-11 Thread Coders
On Mon, 10 Jan 2005 15:02:35 + Dave wrote: DS> > I thought comment changes were always fair game. And comments in DS> > mib2c conf files seems even safer. No? DS> DS> The mib2c conf file are less vital, agreed. But it's not true DS> that comment changes are fair game. What I'd suggested was

Re: non-caching MfD framework

2005-01-11 Thread Coders
On Mon, 10 Jan 2005 15:02:32 + Dave wrote: DS> > I would probably take the existing ${context}_cache_load() and include it DS> > as${context}_container_load(), since that's really what it is doing. DS> DS> Ummm I'm not quite sure what you mean here. DS> Under what circumstances would 'con

Re: non-caching MfD framework

2005-01-11 Thread Dave Shield
DS> Under what circumstances would 'container_load' be called for a DS> non-caching module? When the agent first starts up, during normal DS> operation, or when? RS> Yes, at startup. Then it would be up to the module to maintain RS> it during runtime. Remember, the idea here, by default, is RS

Re: non-caching MfD framework

2005-01-10 Thread Dave Shield
On Fri, 2005-01-07 at 15:23, Robert Story wrote: > DS> (I was a bit surprised to see that you'd applied this > DS> to the 5-2-patches line as well, since I thought that was > DS> currently in pre-release freeze, but that's by-the-by). > > I thought comment changes were always fair game. And comm

Re: non-caching MfD framework

2005-01-10 Thread Dave Shield
On Fri, 2005-01-07 at 15:23, Robert Story wrote: > DS> I'm appending a new "mfd-access" file, that could be used > DS> to generate code which omits all of the cache-related > DS> processing altogether... > What I probably would have done would have been to define a new variable to > make the cache

Re: non-caching MfD framework

2005-01-07 Thread Coders
On Fri, 07 Jan 2005 13:31:15 + Dave wrote: DS> I noticed a log message on the CVS list yesterday, DS> about documenting how to implement a non-caching version DS> of a container-based MfD module. DS> (I was a bit surprised to see that you'd applied this DS> to the 5-2-patches line as wel