1. Eu to C translator questions.
------=_NextPart_000_002B_01BFD983.5AAB9280
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Hi all.
With the Eu --> C translator progressing, I'm curious as to
just how this is all going to work (others have tried to explain=20
it to me but I still don't get it)
How can the Euphoria specific parts of the language (and I'm=20
thinking here of the "dynamic" parts of Euphoria, atoms,=20
sequences etc) be translated to C code without giving the=20
whole game away? Or am right in thinking that an interpreted=20
/ run time core will still exist somewhere along the line.=20
I can't see how the specific Euphoria types can be handled in
C without the code that handles it either being visible to us
or hidden in some sort of run time form.
Can someone explain?
All the best
Mark =20
------=_NextPart_000_002B_01BFD983.5AAB9280
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META content=3D"text/html; charset=3Diso-8859-1" =
http-equiv=3DContent-Type>
<META content=3D"MSHTML 5.00.2314.1000" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT size=3D2>Hi all.</FONT></DIV>
<DIV> </DIV>
<DIV><FONT size=3D2>With the Eu --> C translator progressing, I'm =
curious as=20
to</FONT></DIV>
<DIV><FONT size=3D2>just how this is all going to work (others have =
tried to=20
</FONT><FONT size=3D2>explain </FONT></DIV>
<DIV><FONT size=3D2>it to me but I still don't get it)</FONT></DIV>
<DIV> </DIV>
<DIV><FONT size=3D2>How can the Euphoria specific parts of the language=20
</FONT><FONT size=3D2>(and I'm </FONT></DIV>
<DIV><FONT size=3D2>thinking here of the "dynamic" parts of=20
Euphoria, </FONT><FONT size=3D2>atoms, </FONT></DIV>
<DIV><FONT size=3D2>sequences etc) be translated to C code </FONT><FONT=20
size=3D2>without giving the </FONT></DIV>
<DIV><FONT size=3D2>whole game away?</FONT> <FONT size=3D2>Or=20
am<EM> </EM>right in </FONT><FONT size=3D2>thinking that an =
interpreted=20
</FONT></DIV>
<DIV><FONT size=3D2>/ run time core will still </FONT><FONT =
size=3D2>exist somewhere=20
along the line. </FONT></DIV>
<DIV><FONT size=3D2></FONT> </DIV>
<DIV><FONT size=3D2>I can't see how the specific Euphoria types can be =
handled=20
in</FONT></DIV>
<DIV><FONT size=3D2>C without the code that handles it either being =
visible to=20
us</FONT></DIV>
<DIV><FONT size=3D2>or hidden in some sort of run time =
form.</FONT></DIV>
<DIV><FONT size=3D2></FONT> </DIV>
<DIV><FONT size=3D2>Can someone explain?</FONT></DIV>
<DIV><FONT size=3D2></FONT> </DIV>
<DIV><FONT size=3D2>All the best</FONT></DIV>
<DIV><FONT size=3D2></FONT> </DIV>
------=_NextPart_000_002B_01BFD983.5AAB9280--
2. Re: Eu to C translator questions.
Mark Brown writes:
> How can the Euphoria specific parts of the language (and I'm
> thinking here of the "dynamic" parts of Euphoria, atoms,
> sequences etc) be translated to C code without giving the
> whole game away? Or am right in thinking that an interpreted
> / run time core will still exist somewhere along the line.
> I can't see how the specific Euphoria types can be handled in
> C without the code that handles it either being visible to us
> or hidden in some sort of run time form.
The compiler generates a .c file for each Euphoria file in
your program, i.e. a .c for the main file, and a .c file for each
included file. Creating one huge .c would risk blowing the
size limits on some C compilers.
Your .c's are compiled into .obj's (machine code)
and linked with other .obj's supplied by RDS.
The RDS .obj's contain run-time routines
for things like repeat(), append(), dir() etc. Things which take
too much code to be done in-line.
The scanner, parser, and these runtime routines are shared
between the interpreter and the compiler (although some run-time
routines for the compiler have checking eliminated for
greater speed.)
It's true that by examining the generated C code, people
will learn a lot. They won't be able to examine the run-time
routines because I don't plan to give away the source
to the run-time routines.
Just because people get a few tips on how to implement
Euphoria in C, I don't think that means that a lot of Euphoria
clones are going to suddenly appear, putting RDS out of business.
I don't think there are many people out there with the motivation,
the time, and the stamina to do a really good job of it.
Regards,
Rob Craig
Rapid Deployment Software
http://www.RapidEuphoria.com
3. Re: Eu to C translator questions.
You know I gave this Euphoria compiler project some thought, and imagined
how *I* would do it if I had to. And I came up with a simple sheme. A single
.obj file called Euphoria.obj would contain all the Ephoria run-time
routines.
Then, for each .e or .ew file, a .h header with exported globals is
generated, and a coresponding .c file. Some of the routines in Euphoria.obj
would be something like these;
Sequence * create_sequence();
void init_sequence();
void append_seqeunce();
void prepend_sequence();
void insert_item();
void free_sequence();
Then when a Euphoria program converted to C should use a sequence, it just
does it by calling above routines. Anyone that looks at the source generated
by the Euphoria compiler can't tell how sequence work, because these
functions are defined in a precompiled .obj file.
When thinking about using precompiled routines to implement Euphoria's
library, I came to realise that the mayority or work that needs to be done
by any Euphoria program, is call some C routines. I then thought of how easy
it would be to go for the gold and implement a Euphoria to ASM compiler.
Just some control-flow and math operations, and calling some functions, is
what the ASM source should look like. Look at NASM, it supports dozens of
platforms, consists out off a single executable, is open-source and free,
and works with 32-Bit windows and can call routines from DLLs plus generate
DLLs. I found writing ASM programs in NASM often easier then writing
equivalent programs in C. 32-Bit ASM on Windows is a heck of a lot easier
and more elegant then old-school ASM, infact I find ASM on Win32 to be
easier to use then most "high-level" programming languages out there. (like
JAVA for example : )
Since my guesses are that what Rob is doing is stepping through an array in
memory containing the virtual machine instructions Euphoria programs get
turned into before running them, and then are now translated to equivalent C
code, then mapping the Interpetter's Eu compiled code to C would be harder
then when mapping it to ASM. You should be carefull to nottice if what you
are doing is translate some pointers to "add(),div(),call()" virtual machine
codes to C, because if so the C compiler won't be able to do math
optimisations, and converting output to an assembler instead of a C compiler
would be easier to do that way.
But I guess you allready knew this :)
Jason Leit,
Cheers :)
>Robert Craig writes:
>The compiler generates a .c file for each Euphoria file in
>your program, i.e. a .c for the main file, and a .c file for each
>included file. Creating one huge .c would risk blowing the
>size limits on some C compilers.
>
>Your .c's are compiled into .obj's (machine code)
>and linked with other .obj's supplied by RDS.
>The RDS .obj's contain run-time routines
>for things like repeat(), append(), dir() etc. Things which take
>too much code to be done in-line.
>
>The scanner, parser, and these runtime routines are shared
>between the interpreter and the compiler (although some run-time
>routines for the compiler have checking eliminated for
>greater speed.)
>
>It's true that by examining the generated C code, people
>will learn a lot. They won't be able to examine the run-time
>routines because I don't plan to give away the source
>to the run-time routines.
>
>Just because people get a few tips on how to implement
>Euphoria in C, I don't think that means that a lot of Euphoria
>clones are going to suddenly appear, putting RDS out of business.
>I don't think there are many people out there with the motivation,
>the time, and the stamina to do a really good job of it.
>
>Regards,
> Rob Craig
> Rapid Deployment Software
> http://www.RapidEuphoria.com
________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com
