On Fri, 23 Apr 1999, Jim Schram wrote:
> At 5:28 PM -0700 1999/04/23, Kenneth Albanowski wrote:
> >...My point is to ask whether the *Prv.h headers _are_
> >intended to be useful for developers. If so, then all referenced
> >structures need to be available. (And maybe the header oughn't to be
> >called *Prv.h if it isn't.) If not, why are the *Prv.h headers available
> >at all?
>
> To help you debug... and that's about all, really.
To help debugging by letting you symbolically view the contents of system
structures? Obviously not... To debug by getting _some_ idea of what the
system is doing, certainly. If this structure problem is common (I've not
made a survey), the headers are of more use to people then the compiler.
> Private headers will never be considered 'open for use' by developers.
> They're private...
Well, they're the _publically_ private headers, at any rate. :-)
> that's the point. Thus we reserve the right to change
> anything and everything in there, at any time, without warning. Because
> doing so won't break our published, supported APIs.
No argument.
> BTW, there may very well be cases where the public APIs expose a
> dependency on a private structure... if something like that turns up I'm
> sure we'll take some action to correct the situation and keep the public
> API working for as long as we can (for example, moving the structure
> into a public file... or something like that). So unless otherwise
> noted, any *Prv.* file is ours alone to play with.
>
> Heh heh heh... "Look, But Don't Touch."
Might I suggest some sort of #ifdef, saying roughly:
#ifndef PALM_INC_IS_BUILDING_THE_OS
#error Sorry, building with Prv headers is not supported
#else
//...
#endif
Of course, as David said, I'd worry about getting the public API finished
first. Please excuse my nitpicking. :-)
--
Kenneth Albanowski ([EMAIL PROTECTED], CIS: 70705,126)
