Re: ///Re: NASM problem

På Mon, 08 Jan 2007 05:19:19 +0100, skrev rhyde@xxxxxxxxxx <rhyde@xxxxxxxxxx>:


//\\o//\\annabee <Free" wrote:

In 120000 lines of code, I "needed" this once.

Yes, but we've seen your coding style. I'm surprised you needed it at
all.

Give us an example of a coding style where it will be so important that it would be a significant advantage to have this, then.

That time I replaced it by
3 insignificant instructions.

In this one case. Bottom line is that if you can do something at
compile time, why do it at run-time? We're not talking about the need
to shift an address four bits here, we're talking about the more
general ability to do computations at compile-time rather than at run
time.

Compiletime calculations, in my (admittely limited) experience are needed mostly for teaching application. Tenliners, and such, that is for demonstrations of features noone really actually needs.

You cannot ask such a feature to generate all the CPU instructions, for use in an assembler application, can you? Van you make it generate all the winos structures, and the winos equates? Maybe you can make a compiletime lookuptable with it for a few cases that would be real and needed, but in practice the real target areas of such a thing is few. I have never needed it. I could use it in ONE place in all the code I have written. If it would be something I needed todo each week however, I would view the thing much diffrently. As I have said before, writing a small lookuptable by hand, once, is faster then it would be to study how to make it generated automatic. Some features appear wonderful, but only to the "once who cant do, teach" crowd of people. Also, the compiler would have to generate it each time, slowing down compilation. So this would have also a dramatic effect if used heavily in a huge project, no?

As I am sure that the code I wrote belongs to just a small universe of code, I take reservation that this might be needed in other applications, but I cannot think of any.

Betov said once (if I recall correctly) he was impressed by a guy able to produce a bitmap out of a macro. Yes, maybe was an impressive demonstration, but of what? 180 IQ of pure stupidity?! Nothing noone will ever need for any real thing whatsoever. A demo of "how clever I am"? Plain ludacris! Maybe was a creative stroke of genious, but seems to me useless. Why on earth generate a bitmap in the data or codesegement? Whatfor? Unless I misread something, such macros are waste of time. However, the "call", "push", "Argument" "Proc" and "EndP" and such macros in the RosAsm stantard macros are what I would call real significant and very useful. Thats a whole diffrent world of usefulness.

Not a strong argument in favor.

That's a slippery slope you're headed down. Okay, shifting a
relocatable address isn't very important. Next, doing relative
arithmetic at compile time isn't important. Then, why bother with
constant expressions at compile time?

Because its a much more useful feature. Much more common.

With each step, you're not only
giving up the advantages of assembly language over HLLs, but you're
also giving up features that HLL compilers *can* take advantage of.

They _rarly_ if ever do. See how many such applications you find in the sourcecode of Delphi?

You
talk about *me* not supporting assembly language around here, and then
you come out of left field with Rene and make statements about things
where assembly still has a small advantage over assembly. With assembly
supporters like you, who needs people like Herbert around to bash
assembly?

I dont think Herbert "bashes" assembly. He just hasnt ventured to take on RosAsm. Its not like I dont understand the resistance, I too did not ever thought that something like RosAsm was possible, until I took a real look. The only problem is how to inspire them to take that real good look. RosAsm is seriously addictive when grasped. Seeing the ease of using Herberts dissembler, and how easy it was to compile a com.file with it, I have no doubt he could create just as good an assembler as RosAsm, if he only could see what it would mean. RosAsm is a gigantic step for assembly. The first assembler that will be able to produce applications worthy of competing with current HLL produced applications.

Details are
important, but you are a machine of generalizations of small points that
have no practical serious implications at all.

Until you start to put them together and see that by giving up all
these small little details you give up the power of assembly over HLLs.

No. I do not. I keep bragging about the new editor I wrote, but I cant help it. The time it took for this editor to pass through my fingers, wore so quick, and so painless, and the result such a strong proof of what a beginner assembler programmer can do to blast the current texteditors to bits and pieces, that it will be warming me for years to come.

I tried to write one in Delphi. I think I spent something like several month with it. And still it had so many problems it was better left in the dustbin. This one, written in RosAsm was useful after 3 days! I couldnt belive it myself if I didnt see it with my own eyes!

Let the lord of spagetti and code be my witness, if you know what "beeing in the zone" could mean, there is _nothing_ on earth comparable to beeing in the RosAsm zone. Better then meatbolls and potatos, better then french wine, and in fact, also a strong competion to "chasing arse", like Hutch, the old fart, would have said. Of couse, i dont _have_ to chase ars, and there in ly his confusion.

MASMTOFF, Hutch! My editor makes "the gun" look like a pile of childish blood and tears !



Cheers,
Randy Hyde


.

Relevant Pages

  • Re: ASM vs HLL : absurd war
    ... assembler portion of RosAsm only constitutes 8,100 machine instructions. ... that this is due to programming style rather than language syntax. ... admit that you don't know any HLLs other than COBOL, ...
    (alt.lang.asm)
  • Re: This Crashes RosAsm
    ... > RosAsm can, of course and evidently hold any number ... > of the actual Assembler. ... > labels in a very effective manner. ... > * In case MASM would compile such a list of Labels ...
    (alt.lang.asm)
  • Re: Interesting Web Site on Open Source Development
    ... why is RosAsm so broken? ... going to be a simple fix, they just move on to a different assembler. ... >> is faster than FASM by comparing the compile times of RosAsm ... Comparing yourself with MASM? ...
    (alt.lang.asm)
  • Re: The HLL Temptations
    ... >> structured design, scalability, lifespan, suitability of the language ... People are dropping their HLLs and switching to use assembly ... Oh well, while RosAsm is floundering, other assemblers *are* adding ... >> I've programmed all my adult life, from assembler to ...
    (alt.lang.asm)
  • Re: ADK/HLA v2.0 Update
    ... The "ideas and systems" we implement in RosAsm are ... >> have implemented the Wizards and some Aplications Builders, ... >> come to RosAsm, but some of the guys using some HLLs (not C, ...
    (alt.lang.asm)