From: E. Gladyshev (egladysh_at_[hidden])
Date: 2003-09-17 11:24:49
--- Michael van der Westhuizen <Michael.vanderWesthuizen_at_[hidden]> wrote:
> Nice work.
> I'm a little concerned that a lot of the concepts
> that are being wrapped up (notus/gdi/bitmap.hpp) are
> not portable concepts. The idea of a HBITMAP being
> wrapped up with detail namespaces and typedefs etc.
> for source compatibility is very nice... but... I
> have my doubts about how portable it is (X, Carbon).
I am working on the wxWindows support right now.
It is looking pretty good so far.
I think that we'll be able to release this sample
for wxWindows rather quickly.
If we have wx, then we'll have Mac/Unix automatically.
Actually I was a bit suprised how similar wx is to win32.
As for HBITMAP
typedef HBITMAP hnd;
typedef wxBitmap* hnd;
Also note, one can always create custom types in the
details space to accomodate the policies specs.
for a particular platform.
> I think a lot of rework would be avoided by doing
> the initial proof of concept work in parallel on
> Win32, X and Mac. These three platforms are so
> radically different any Win32 conceptual hangovers
> would be identified early.
I agree. This is why we want to put wx into the picture
at the early stage. It'll help us to refine the
implementation traits/policies specification
for a better portability.
> The design docs look good. What does remain to be
> seen is how message despatching (X event loop,
> Win32 API message handlers, Mac??) works across
Amazingly, wx has the same concept of the message loop
as win32 does. Since wx is portable acrorss Max and
other platforms, they managed to provide the message
loop support for all the other platfroms.
So I think we should be ok here.
Do you Yahoo!?
Yahoo! SiteBuilder - Free, easy-to-use web site design software
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk