Re: i disagree



David Jones <ncic@xxxxxxxxxx> écrivait
news:MPG.202477782f78f69c9896bb@xxxxxxxxxxxxxxxxxxxxx:

OK, whatever. s/assembler/compiler if it makes you happy. The
argument
still stands: if doing just text processing (ie: macros) maintains a
1:1
relationship between source and emitted machine instructions, then
would
you consider such a compiler 1:1?

Yes, of course.


Isn't that absurd? What _couldn't_ be compiled through text processing
alone before the final pass? And how is this any different from a
user-
defined macro?

This is exactely where one of the Randall Hyde swindlings
stands, playing with the evident fact that most of you fail
to understand that the dilemna is not with a *what*, but
with a *how*.

Again, the very same Statement, found in a Source, may be
either HLL or Assembly depending on the way it is processed,
and not at all on what it is.

By definition, a Macro-Assembler is an Assembler coming with
a Macros System, that enables with HLLisms, as a writing Style,
and, by definition, the Macros belong to the Source and are
under the control of the programmer.

Therefore a Macro can, in no way break the 1:1 correspondance,
as long as they are nothing but a way of writing a Source,
would the Macro expand into 1 or 248 Instruction(s)... written
by the Programmer.

Now, as soon as, *inside an Assembler*, an Assembler's Author
nests hidden Macros (because it is way easier to implement it
that way, than to develop a powerful enough macros System
-see MASM Proc, for example-), the name is no more *Assembler*,
but *Compiler for Assembly Language*: When the programmer cannot
know what he is doing, unless making use of a Disassembler, this
is the final demonstration that he has NOT been writting Assembly.


Betov.

< http://rosasm.org >




.