Re: Other anticipated changes for 0.8.5, Part 2 Mark S. Miller (markm@caplet.com)
Tue, 26 Oct 1999 11:06:04 -0700

At 06:07 AM 10/26/99 , Brian Marick wrote:
>A bunch of those Trace objects are (were) created when the TraceController
>is initialized, but any class/subsystem/code can create additional ones. So
>there doesn't have to be a single central chunk of code where all Trace
>objects are created. The predefined ones are there only for convenience,
>not out of necessity.
>
>Or do I miss your drift?

You got my drift exactly. I, perhaps, misunderstood the history. I thought we were deprecating the definition of Trace objects by classes outside the trace subsystem. I thought non-legacy code was supposed to instead use only Trace objects exported by the trace subsystem.

>I'm on the road, so I can't look at the code right now. But I think that's
>how it worked.

That is how the code works. Both styles are fully supported. The question is one of intent. If you didn't intend to deprecate decentralized Trace-object definition, then I evolved the code towards the wrong consistency. Instead, as I think I prefer anyway, we should migrate each Trace-object definitions that is not *about* the trace subsystem into the subsystem it is about. 

         Cheers,
         --MarkM