Re: This Crashes RosAsm

From: Randall Hyde (randyhyde_at_earthlink.net)
Date: 07/01/04 

Date: Thu, 01 Jul 2004 14:05:12 GMT

"Betov" <betov@free.fr> wrote in message
news:XnF951967D89A81Fbetovfreefr@212.27.42.66...
> arargh406NOSPAM@NOW.AT.arargh.com écrivait
> news:gdf6e05mst0hrr17dtqms4k7b6r5814a2v@4ax.com:
>
> > ANY production grade assembler
> > or compiler should be able to handle ANY number of unique symbols
> > without breaking, up to the point where the symbol table overflows,
> > and at that point, a reasonable error message should be issued.
> >
>
> RosAsm can, of course and evidently hold any number
> of Symbols.

Evidently. Except, of course, for the sample program
I've posted to this newsgroup. I can provide several
others, if you'd like.

>
> For the ones who could given Master-Self-Esteemed
> Randall Hyde stupidities some credit, here is some
> explanations about what is going on:
>
> * Master Pdf is much dispited with having to face
> a release of RosAsm that makes it, now, the fastest
> of the actual Assembler.

Fastest is meaningless if it doesn't work.
For all intents an purposes, RosAsm is currently
the *slowest* assembler because all the other
assemblers I've tried *can* assemble the test
file I've provided here. RosAsm crashes and
never completes the job (effectively taking an
infinite amount of time to assemble that short file).

>
> * RosAsm was already, before this recent 3x speed-up,
> one of the fastest Assemblers, aside FASM. At that
> time, Master Pdf, run into some vile and unhonest
> tests, that he made, _on purpose_, weird, to try to
> show that ... MASM (???!!!...) was faster than RosAsm.
> He achieved this great performance, by playing on the
> fact that RosAsm was not designed to hold synthetic
> labels in a very effective manner. So, of course, he
> _got_ his "Demonstration", and we have seen every
> here and there, great debate where this *** was
> making a great noise and claiming that RosAsm was
> ... _slow_ (???!!!...).

Slow isn't the issue right now, Rene.
Correctness is the issues. RosAsm simply doesn't work
on my test file. Slow, people can handle. Crashing....

>
> * The most amaizing thing, to me, in this kind of
> affair, is that there are still some people giving
> some credit to this swindler. This shows clearly
> at least one thing: The mythology of "teacher",
> "books", "Expert talking from the top of his teacher
> seat", is still very active, and a complete lack of
> any kind of competency, in programming, can easily
> be enough if you use pedantic and scolastic words,
> and if you claim that you are an expert, even without
> any base... :(

Yes, it's somewhat amazing that some people expect
a program to *work*. As you've no doubt read in
many posts around here, most people don't care if
RosAsm self-compiles itself in two or six seconds.
Both time frames are fine for that magnitude of a
program. But the fact that RosAsm crashes on
a short program with fewer than 1,000 symbols
is a big issue to a lot of people. At some point,
you'll need to figure that out.

>
> * Now, what is happening is nothing but what already
> happend previously. _Exactely_. Now that Master Pdf
> can no more play his vile and durty bad trick with the
> synthetic Symbolics thingie, he is trying to do it again
> with... a list of Labels (???!!!...). The final goal of
> it, is to compile a List of Labels with MASM and with
> RosAsm, and to compare the timings (???!!!...).

Yes, the final goal of this list of symbols is to verify
your claims that RosAsm is now the fastest of all
assemblers. Do you have a problem with independent
verification of your claims? I suspect that this is the
main reason you refuse to fix this defect: because you're
afraid your statements will be shot down. But keep in
mind that people are *far* more willing to accept
RosAsm as "not quite the fastest assembler, but still
very fast" rather than "it crashes on simple input files."

>
> * In case MASM would compile such a list of Labels
> faster than RosAsm (i have really no idea, but this
> seems to me perfectely possible...), we would see,
> again, tons of posts from this swindler claiming that
> he has _proven_ that MASM is faster than RosAsm...
> and, again, the most amaizing thing is that there
> _will_ be some clients to buy his bullshits...

Never can tell. MASM might actually be *slower*
at this. Until you fix your problems, we cannot verify
this, now, can we? I would ask why it's okay for you
to make claims like "this is the fastest assembler" without
any proof of that statement, yet I'm a "swindler" when I
attempt to run some tests to verify your statement.
Fix the bug. Then we'll know once and for all what the
truth is. If you can't handle the truth, then you've got
a lot of nerve calling me a "swindler." (not that you
haven't demonstrated a lot of nerve in the past.)

>
> * Actually, the fact is that RosAsm has never been
> written for compiling a dictionary, and i have not
> even _one_ reason for ever implementing an error
> Message to hold those cases that will _never_ be
> seen in any Source of any beginner trying to write
> any Application.

This isn't a dictionary. It's a set of labels that were
taken from a dictionary because *you* complained
about the use of synthetic labels in past benchmarks.
Would you prefer I go back to using synthetic labels?
And if the label "actual" (in combination with other labels)
is a problem for your assembler, shouldn't this be a
concern to you? Was RosAsm designed not to handle
labels like "actual"?

>
> * So, why should i impement an error Message, that
> nobody -but Master Self-Esteemed Randall Hyde- could
> ever see? Simple: To enable Master Self-Esteemed
> Randall Hyde with going to an end with his famous
> "BenchMark"... and with runing, again, into his
> famous propaganda threads selling incredibles lies
> and bullshits on young head... :)

So no one else will ever use a combination of symbols,
including "actual" that will trip this defect in your code?
Cheers,
Randy Hyde

Quantcast