OSLib / wimp.h / wimp_ERROR_BOX_CATEGORY

[Alexander Thoukydides]

In <URL:news:local.oslibchat> on Sun 04 Jun, Tony van der Hoff wrote:
> On Sat, 3 Jun 2000, at 11:45:15, Tom Hughes <tom at compton.nu> wrote on
> the subject "OSLib / wimp.h / wimp_ERROR_BOX_CATEGORY":
> >In message <QCttIGAADQN5EwXv at mk-net.demon.co.uk>
> >          Tony van der Hoff <OSLib at mk-net.demon.co.uk> wrote:
> > > On Wed, 31 May 2000, at 10:34:35, Dave Appleby
> > > <DaveAppleby at btinternet.com> wrote on the subject "OSLib / wimp.h /
> > > wimp_ERROR_BOX_CATEGORY":
> > >
> > > > I think that using the shift value for the new categories makes
> > > > things unnescessarily complex. Surely the programmer doesn't need to
> > > > know how the flags are arrived at...
[...]
> > My definitions were merely following Jonathan's example set in other
> > parts of OSLib for how to deal with the case where a bit mask includes a
> > set of bits used to hold a small ordinal value.
> >
> Well, in that case, while I have some sympathy for the original
> suggestion, I'd be disinclined to follow it up right now, on the basis
> that to be consistent, we'd have to make similar amendments throughout
> the library. The present system has been considered adequate for years,
> AFAIK.
[...]
> The lack of comment suggests that the general consensus of oslib-user isn't
> strongly either way?

My personal preference would be to have both sets of values. The final bit
patterns are easier to use when constructing or modifying flags, but the
enumerated value together with a shift is more useful when decoding an
existing field.

However, if only one set of values are to be provided then the enumerated
value combined with a shift is the most useful.
-- 
--  Alexander Thoukydides  Email: mailto:alex at thouky.co.uk
--  Bournemouth, England     WWW: http://www.thouky.co.uk/
... Microsoft is not the answer.
... Microsoft is the question.
... NO (or EPOC) is the answer.