OD>as well.. (I guess.. Which is first in the far method: segment or
OD>offset?)
TH>Segment is adx+1.
OD>Huh? What is ADX? It is the 32bit version of DX, as EAX is of AX?
TH>Sorry, "adx" is short for the address of the instruction data, and
OD>Address of the instruction data?  Please elaborate...
Well, the effective addresses for instructions requiring a far
address are always in the form [segment:offset]. The offset is the
first two butes and the segment follows. Instructions that come to
mind are the LEA, LDS, LES, and the various FAR jumps and calls.
These instructions get their data either as an imediate operand, or
are referenced from some other location, ie. indirect. In either case
it is the instructions data, ie. the pointer.
TH>oops, I ment adx+2. My point was that the segment of the address is
TH>two bytes higher than the offset. And, ofcourse, I am refering to the
TH>16-bit model.
OD>adx+2 ....because both are expressed as Words?
They are words, yes. And, the offset is in bytes, hence '+2.'
OD>"not taken a speed hit, yet"... meaning your routines are fast????
TH>On a 586 they sure are. But, what I ment was that compile times are
TH>only two or three seconds for even those file sizes. PowerBASIC is a
OD>Yup. My Quick-Basic compiler happens to be fast, too. The largest
OD>I compiled must be my SamsalP demo... Didn't take THAT long to
OD>compile...
I used QuickBasic before. I don't think I know the internals of
either to opin on which one is better overall. PowerBASIC has some
nice built in features QuickBASIC doesn't have. PB also has some
irritating bugs.
I gusee what I like about PowerBASIC is that it comes with a command
line compiler. I like that, because PowerBASIC, like QuickBASIC, come
with crummy IDEs, IMHO. I wrote an IDE much more to my liking.
TH>true compiler and produces machine code. As for how effient it is,
TH>I have not performed any analysis.
OD>What, there are "fake" compilers?
Well, of course there are no 'fake' compilers. [grin]
But, some compile 'P-code' files which are not executable. They are
tokens and data that an interpreter can follow more effiently than
the old interpreters of the old days like GWbasic and IBMs basic, but
are not as fast as true compilations. That's what I ment.
cya
--- Maximus/2 3.01
---------------