Re: Getting NASM from C

From: Beth (BethStone21_at_hotmail.NOSPICEDHAM.com)
Date: 06/01/04 

Date: Tue, 1 Jun 2004 03:42:13 +0100

Frankie say:
> assemblyperson wrote:
> > I appologize for posting this as a reply to my original
post. My news
> > server is not giving me the current state of the news group.
>
> Y'know, we had a "flap" a while back (I kept out of it) that
involved
> posting times and time zones and stuff. On top of that, news
servers are
> erratic - it's not unusual for a post with a two or three day
old date
> to pop up as "new". I know it's not a "global" problem - in
some cases,
> I see a reply a day or two before the message - comcast's
fault, in that
> case, I guess... I wondered if we were "talking past each
other" here...

I think that might be involved...plus, yes, I was probably still
"over-defensive" from elsewhere that I wrongly was still acting
that way in reading assemblyperson's post...took it all the way
when he meant nothing of the sort...

As I note Wannabee made a "don't talk in the third person"
comment (and knowing what he's like in hounding me around ;)
then he probably has also noted that I've also been taking on a
whole bunch of Microsoft "fans" over on another forum in
SourceForge...and that has also contributed here, even though
it, indeed, has nothing to do with this forum that it shouldn't
have "spilled over" from there...if you're wondering, then one
of the posters replied to me using the third person throughout
and I complained that this was a potentially rude thing to be
doing...might be a coincidence only...but knowing Wannabee - my
*ahem* number one "fan", unfortunately - he saw that and that's
why he made the fuss about "using third person is very rude"
comment in that other post...

So, there's "talking passed each other" but also a touch of
"talking to someone else" too...you know, writing one post where
you're really annoyed and, oops, accidentally taking over that
"bad mood" from one post to the other...I, unfortunately, do
have a tendency to sometimes do that because I don't really
realise just how foul a mood I'm in at the time and wander over
straight away to another post without "taking a breather" to
calm down again...sorry...a purely unintentional mistake that I
have a terrible habit of repeating every once in a while...

> ...
> > In fact, I expected that Beth would be an enthusiastic
contributor to
> > the new found page. It seems to solve
> > all of the problems and takes people up on their many
constructive
> > suggestions. I had to do a few things
> > myself to make this happen. I hope Beth feels better about
it. Maybe
> > someday I will have permission to speak
> > to her.
>
> I've never known Beth to refuse to speak to anyone, no matter
*how*
> pissed she is. She prides herself on being fair-minded and
forgiving. I
> think you'll be fine... give it a little time...

Oh, totally...I screwed up here...misread assemblyperson's
post...was still pissed off about a post to someone completely
different that, indeed, had NOTHING to do with assemblyperson
but, well, "wrong place, wrong time" that, oops, my "bad mood"
spilled over from one post to another...

Indeed, _THAT_ was what made me misread what I was seeing...made
me automatically presume that it was some kind of attacking
sarcastic comment or something...when, on the contrary, it was a
great thing and now I've got things straight again, is good and
I'd be _encouraging_...totally great idea, in fact, there, if
you've got some webspace for it all...Annie's also got a point
that CVS isn't perhaps the best place to put it (but, well, I
didn't have anywhere else to put it...plus, I was thinking "this
will be useful to LuxAsm"...although, you know what? Now with
all this talk of XML and perhaps doing it in a more "generic"
manner, it's sounding a lot more like it's "ConvInc" rather than
"LuxAsm" I should be popping it into, anyway ;)...

No, indeed...NO-ONE needs "permission" to speak to me...the sole
requirement for my co-operation is a modicum of, you know,
polite, civil and honest communication...yeah, even Annie just
marginally qualifies for that too...though, she does skirt the
borders at times that, indeed, perhaps the sole person who I
would actually consider "blanking" and "sending to Coventry", as
we like to say over in these parts...

> > People wishing to join that other effort were cautioned
(perhaps too
> > threateningly) that signing its required
> > non-disclosure or secrecy agreement would indeed be binding
and would
> > not be something that they should take
> > lightly.
>
> Okay, that's probably not something I'm interested in (never
can
> tell...)(boom, boom). So long as this syscall documentation is
of the
> "information wants to be free" persuasion, we can cooperate on
that.

Ditto...

> > I got permission to put up information on a certain subject
at
> >
> > http://scienceholdings.com/linux-nasm/
>
> Great! As you know, "disappearing web sites" are an ongoing
problem.
>
> > The permission is not contingent upon using a particular
format.
>
> Seems to me this "format issue" has stymied projects before -
the
> "convert AoA16 to Nasmese" project fizzled out before we'd
decided what
> format we were going to edit it in... AFAIK, the reason for
choosing the
> "RBIL format" was that tools exist to translate it to other
formats. The
> "content" is already separate from the "presentation". The
idea being
> that we edit the "source" document - the RBIL format - and
when it's
> done, run it through the "RBIL to html" tool (I haven't
actually tried
> this with Beth's file...), and have something ready to put up
on your
> site, or whatever.

Yup, you've got the basic idea there, Frank...and, yeah,
"formats" and such _are_ often capable of killing some projects
stone dead because no-one can even agree a basic format so
nothing can proceed...

Indeed, you probably appreciate, therefore, why I was stressing
and kicking up a small fuss about talking and communicating and
agreeing "formats" and "standards" and so on and so forth
_BEFORE_ beginning work proper...yes, it slows things down but
_for good reason_ in that if you can solve these things at the
beginning then you can Hopefully guarantee that everyone is "on
the same page" throughout a project...because, in a sense, this
also kills off projects half-way through too when work actually
commences and people start disagreeing about what work has been
done and then it all "fragments" in different directions...

> Maybe XML is "better" at this. I don't know squat about XML.
I'll check
> out those links Jim posted (thanks, Jim!). I see "XML" and
"XHTML" in
> your document, but then it seems to be plain html... Is that
all XML is?
> Even if XML is better, it won't hurt to do it in "RBIL
format", or *any*
> format consistent enough to be "machine readable", first - it
can be
> converted to XML (or whatever format) from there, no?

Well, HTML is a specific "subset" of XML for displaying
webpages...

XML in itself is, in fact, so incredibly easy that you'll likely
be thinking "huh?!? Is that _all_ it is? Isn't there something
more complex?"...like HTML, it's a text file with "tags" inside
it...and then there's an opening and closing tag (e.g. "<tag>"
and "</tag>" :)...

And would you believe it? Other than it's much stricter about
things than HTML (every open tag _must_ have a close tag...so
"<p><br><p>" would be an error in XML, though a HTML browser is
written to "tolerate" such "errors"...although, there is the
"<tag/>" form which is kind of an open and close tag all in
one...so, "<br/>" would be the proper XML way to handle this tag
which has no close equivalent :)...

Other than that, this _IS_ XML...that's it...what tags does it
support? Well, that's kind of the point...they are all
"user-defined"...you just make up your tags as you go...and
that's really what HTML is, as a (less strict) "subset of XML",
in a sense...they've just defined specific HTML tags with known
meanings for a web browser to display it...

For example, this could be one way to store the syscall list in
XML:

---------------------------------------------

<?xml version="1.0" encoding="ISO-8859-1"?>

<LIST>

   <NAME>Linux/x86 SysCalls</NAME>
   <DATE>
     <DAY>1</DAY>
     <MONTH>6</MONTH>
     <YEAR>2004</YEAR>
   </DATE>

   <FUNCTION>
     <NAME>sys_exit</NAME>
     <DESCRIPTION>TERMINATE THE CURRENT PROCESS</DESCRIPTION>
     <CATEGORY>Kernel System Call</CATEGORY>
     <CALL>
       <PARAMETER>
         <NAME>Function number</NAME>
         <REGISTER>EAX</REGISTER>
         <VALUE>000000001h</VALUE>
       </PARAMETER>
       <PARAMETER>
         <NAME>Exit status</NAME>
         <REGISTER>EBX</REGISTER>
       </PARAMETER>
       <INTERRUPT>80h</INTERRUPT>
     </CALL>
     <RETURN>
     </RETURN>
     <ERRORS>
     </ERRORS>
     <NOTES></NOTES>
     <SOURCEFILE>kernel/exit.c</SOURCEFILE>
     <SEEALSO></SEEALSO>
  </FUNCTION>

</LIST>

---------------------------------------------

Note that the first line is merely a "sanity check" that says
"yes, this is a XML file (version 1.0 XML, in fact)" and
specifies the character set...just included to make sure it gets
interpreted correctly...

Otherwise, it's completely "free-form"...just a text file where
you stuff things in between "tags" that mark them as being
"something"...that's it...

The beauty in XML is exactly because it's basically rather
simple and minimal yet flexible and powerful...you can store the
data in any old way you like...

Oh...one other difference to HTML is that XML is case
sensitive...I've got "NAME" above and this isn't the same as
"name"...

But because XML is all about "tags" then it's really easy to
modify...for instance, if I wanted to add some new piece of data
then I just pop into some "<NEWDATA></NEWDATA>" tags in the
right place and fill them up with information...there's also no
strict requirement from XML itself that every "<FUNCTION>"
necessarily has the same types of data inside it...that would
likely be "expected" from the point of view of some program or
whatever parsing the file...but, for example, I've called it
"<FUNCTION>" rather than "<SYSCALL>" and specifically added a
"<CALL>" tag to detail the calling sequence with an explicit
"<INTERRUPT>80h</INTERRUPT>" tag...the idea being that you could
possibly "extend" this format to also have a
"<STDCALL>_ExitProcess@4</STDCALL>" instead, which says "nope,
not an interrupt this time but a STDCALL call function instead
with this name"...

And the whole thing is a simple plain text file...so, really, it
is, indeed, not a bad little idea to use XML...it's
machine-readable, basic, flexible, simple plain text...and
surprisingly free - when taken in "pure XML" form without the
XSL / XSLT stuff or whatever - of the usual complexities that
these kinds of things usually have...I mean, it would be
relatively simple to handle this with a rather uncomplicated
program from DOS or whatever too...no dependency on anything
daft like "Javascript" or whatever...

Just a case of looking for "<" then reading the "tag" until the
">" (like HTML, you can also have "attributes" like: "<INTERRUPT
VECTOR="80h"/>" too, if you really wanted :)...and then matching
up open tags with close tags (but XML is _strict_ unlike HTML,
so it must be properly nested and every open tag matched with a
closed tag or it is an error and processing should stop and the
file "rejected"...which is useful because a lot of complexity in
HTML browsers comes from the "loose" way that it accepts any old
crap you pass to it - full of errors or not - and just renders
_something_...not so in XML that it's got to be "well-formed" or
it's rejected which makes it easier to parse...if what you're
expecting doesn't happen then just bomb out with an error ;)...

> > Kindness, integrity, and courage form the basis of team
success.
>
> Sounds about right to me! I might also add "flexibility"...

I'd also add "communication", from a practical point of
view...plus, there's "compromise" too (but your "flexibility" is
arguably already covering that, just with a different word ;)...

Beth :)