Maybe I'm way off BUT it would be nice IF REALbasic were more than
capable enough to do this without needing C/C++.
I rarely write any assembler in C/C++ any more because it's just not
necessary for 99.999% of the tasks I use C/C++ for.
Compilers are pretty darned good and chunk out pretty decent code such
that it works more than fast enough for the task without resorting to
writing assembler.
For RB, C and C++ are the "assembler" equivalents. and it will be nice
to get an RB compiler with enough optimizations, etc that C/C++ are
mostly not needed in the same way.
I must confess that I don't understand why certain things are sooo much
faster in C than in RB. C will clobber RB in a simple loop even with all the
Pragmas. The difference is huge! Is there some aspect of RB loops (not the
MT because that can be turned off via Pragma) that is slowing them down?
Same with heavy math.
On the other hand, some stuff I understand, but I wish RS would put the
effort into overcoming it. C has always been great with direct memory
manipulation and bitshifting, i.e. arrays of bytes (or chars or longs or
what have you). RB pushes this stuff through methods of classes like
MemoryBlock and Bitwise, where C probably outputs inline code. It would be
nice to have some additions to the language that generate tight, inline code
for certain operations, rather than going through the method dispatch table
of a class.
Having said that, RS feature request list must be huge. As long as they keep
moving forward on their list, they can have my subscription. Having a solid
cross platform tool is that important to me.
Daniel L. Taylor
Taylor Design
Computer Consulting & Software Development
[EMAIL PROTECTED]
www.taylor-design.com
_______________________________________________
Unsubscribe or switch delivery mode:
<http://www.realsoftware.com/support/listmanager/>
Search the archives of this list here:
<http://support.realsoftware.com/listarchives/lists.html>
