Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.forth > #55885 > unrolled thread

Re: bug in VFX --- IF does't work

Back to article view | Back to comp.lang.forth

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: bug in VFX --- IF does't work hughaguilar96@gmail.com - 2017-03-18 00:30 -0700
    Re: bug in VFX --- IF does't work stephen@mpeforth.com (Stephen Pelc) - 2017-03-18 13:15 +0000
      Re: bug in VFX --- IF does't work Jan Coombs <jenfhaomndgfwutc@murmic.plus.com> - 2017-03-18 13:54 +0000
        Re: bug in VFX --- IF does't work hughaguilar96@gmail.com - 2017-03-18 18:45 -0700
      Re: bug in VFX --- IF does't work hughaguilar96@gmail.com - 2017-03-18 13:24 -0700

#55885 — Re: bug in VFX --- IF does't work

On Sunday, June 14, 2015 at 6:13:33 PM UTC-7, Doug Hoffman wrote:
> On 6/14/15 7:48 PM, hughaguilar96@gmail.com wrote:
> > On Sunday, June 14, 2015 at 4:20:31 PM UTC-7, Stephen Pelc wrote:
> 
> >> If you had made the slightest effort, e.g. by disassembling LITERAL
> >> and what it contains, or by reading the manual, you would find that
> >> LITERAL is already  effectively DEFERred.
> 
> > I looked up LITERAL and found that it is not deferred. The manual says: [snip]
> 
> Hugh,
> 
> Stephen also suggested disassembling LITERAL.  Try it.
> 
> -Doug

For anybody who cares (I don't know why anybody would), here is a disassembly of LITERAL under VFX:

see literal 
LITERAL (is IMMEDIATE) 
( 00411474    E8F35DFFFF )            CALL      0040726C        CLIT,
( 00411479    C3 )                    NEXT,
( 6 bytes, 2 instructions )
 ok
hex 0040726c disasm/f 
( 0040726C    E8ADB30000 )            CALL      0041261E
( 00407271    48 )                    DEC       EAX
( 00407272    B543 )                  MOV       CH, 43
( 00407274    00FF )                  ADD       BH, BH
In the weeds ...
( 00407276    0000 )                  ADD       0 [EAX], AL
( C bytes, 5 instructions )
 ok
hex 0041261e disasm/f 
( 0041261E    8D6DFC )                LEA       EBP, [EBP+-04]
( 00412621    895D00 )                MOV       [EBP], EBX
( 00412624    5B )                    POP       EBX
( 00412625    8B13 )                  MOV       EDX, 0 [EBX]
( 00412627    8B5D00 )                MOV       EBX, [EBP]
( 0041262A    8D6D04 )                LEA       EBP, [EBP+04]
( 0041262D    FFD2 )                  CALL      EDX
( 0041262F    C3 )                    NEXT,
( 12 bytes, 8 instructions )
 ok

LITERAL calls CLIT and that is all that it does. CLIT (which is apparently smudged because it can't be found in the dictionary) is located at $0040726c. All it does is call another function ($0041261e), and that function uses the return-address as a pointer, so it doesn't return to its caller (the CLIT function) but instead eventually returns to the caller's caller (the LITERAL function). That function ($0041261e) does a vectored call. It first pushes EBX (the TOS register) onto the data-stack. Then it pops the return-address into EBX. Then it moves the datum at this address (what was after the CALL in the $0040726c function) into EDX. Then it pops the data-stack back into EBX. Then it calls through the EDX pointer. So, the vector is after the CALL instruction in the $0040726c function --- it is located at: $00407271 --- note that we have data juxtaposed with code, which is a bad idea because it results in cache thrashing.

I don't know why the $0041261e function wasn't written like this:
            pop edx
            mov edx, 0 [edx]
            jmp edx

Actually, I don't know why LITERAL wasn't just made a deferred word --- VFX does have deferred words (such as NUMBER? etc.) --- none of this convoluted assembly-language was necessary.

Anyway, none of this matters. It is a waste of time to become an expert at the internal workings of VFX --- there is no guarantee that any of this will still be available in a new release of VFX --- that is the downside of relying on carnal knowledge.

The bottom line is that the Forth-200x committee (which includes Stephen Pelc) will never define LITERAL 2LITERAL and FLITERAL as being vectored words --- they don't want the Forth community to write cross-compilers in standard Forth that use the built-in outer-interpreter and support literal numbers in TARG colon words --- ANS-Forth and Forth-200x are both purposely crippled to prevent this.

[toc] | [next] | [standalone]

#55898

On Sat, 18 Mar 2017 00:30:26 -0700 (PDT), hughaguilar96@gmail.com
wrote:

>For anybody who cares (I don't know why anybody would), here is a disassemb=
>ly of LITERAL under VFX:
>
>see literal=20
>LITERAL (is IMMEDIATE)=20
>( 00411474    E8F35DFFFF )            CALL      0040726C        CLIT,
>( 00411479    C3 )                    NEXT,
>( 6 bytes, 2 instructions )
> ok

What a prat you are! The boss at an engineering company I used to
visit has a sign in his office:
  " Activate brain before engaging mouth"

dis clit, 
CLIT, is DEFERred. Action is C_LIT 

Stephen

-- 
Stephen Pelc, stephenXXX@mpeforth.com
MicroProcessor Engineering Ltd - More Real, Less Time
133 Hill Lane, Southampton SO15 5AF, England
tel: +44 (0)23 8063 1441, fax: +44 (0)23 8033 9691
web: http://www.mpeforth.com - free VFX Forth downloads

[toc] | [prev] | [next] | [standalone]

#55899

On Sat, 18 Mar 2017 13:15:25 GMT
stephen@mpeforth.com (Stephen Pelc) wrote:
> [...]
> What a prat you are! The boss at an engineering company I used
> to visit has a sign in his office:
>   " Activate brain before engaging mouth"

Nicely pruned, thanks, I didn't want to read the rest. 

Jan Coombs

[toc] | [prev] | [next] | [standalone]

#55915

On Saturday, March 18, 2017 at 6:54:53 AM UTC-7, Jan Coombs wrote:
> On Sat, 18 Mar 2017 13:15:25 GMT
> stephen@mpeforth.com (Stephen Pelc) wrote:
> > [...]
> > What a prat you are! The boss at an engineering company I used
> > to visit has a sign in his office:
> >   " Activate brain before engaging mouth"
> 
> Nicely pruned, thanks, I didn't want to read the rest. 
> 
> Jan Coombs

The ANS-Forth standard is crippled to prevent ANS-Forth programmers from writing cross-compilers that use the built-in outer-interpreter and support literal numbers inside of TARG colon words. This was done to prevent them from doing so in competition with the ANS-Forth committee members. Now, Stephen Pelc is on the Forth-200x committee and he is striving to cripple Forth-200x in a similar manner. He wants the Forth community to buy his cross-compiler --- he wants to prevent them from writing their own cross-compilers.

Why are you brown-nosing Stephen Pelc? Do you work for MPE? Did you buy a copy of VFX and now you feel obliged to brown-nose MPE? Public brown-nosing is disgusting! You must have some monetary reason for making such a disgusting spectacle out of yourself. Tell us!

[toc] | [prev] | [next] | [standalone]

#55903

On Saturday, March 18, 2017 at 6:15:20 AM UTC-7, Stephen Pelc wrote:
> On Sat, 18 Mar 2017 00:30:26 -0700 (PDT), hughaguilar96@gmail.com
> wrote:
> 
> >For anybody who cares (I don't know why anybody would), here is a disassemb=
> >ly of LITERAL under VFX:
> >
> >see literal=20
> >LITERAL (is IMMEDIATE)=20
> >( 00411474    E8F35DFFFF )            CALL      0040726C        CLIT,
> >( 00411479    C3 )                    NEXT,
> >( 6 bytes, 2 instructions )
> > ok
> 
> What a prat you are! The boss at an engineering company I used to
> visit has a sign in his office:
>   " Activate brain before engaging mouth"
> 
> dis clit, 
> CLIT, is DEFERred. Action is C_LIT 

The point that I'm making is that ANS-Forth does not vector LITERAL 2LITERAL and FLITERAL --- I think this was done for the purpose of preventing ANS-Forth programmers from writing a cross-compiler that uses the built-in outer-interpreter and supports literal values in TARG colon words --- also, the Forth-200x committee (which you are a member of) refuses to vector LITERAL 2LITERAL and FLITERAL for the same reason.

You are now telling me that CLIT, is a deferred word. I am pointing out that CLIT, is not ANS-Forth. Also, nothing in ANS-Forth is deferred. You are trying to trap VFX users in vendor lock-in. You are telling me that I have to disassemble VFX and figure out how LITERAL works internally in order to write a cross-compiler that uses the built-in outer-interpreter and supports literal values in TARG colon words. That information would only be useful to me if I were going to buy VFX --- but I don't want to be an MPE customer --- I want to program in ANS-Forth. If you are opposed to people programming in standard Forth, but insist that they buy your VFX, then you are the worst qualified person in the world to be on the Forth-200x committee --- you are not supporting the Forth community --- you are giving the Forth community a crippled standard in order to push them into buying VFX.

BTW: FLITERAL is not vectored in VFX --- so even if I pay for VFX I still don't get to write a cross-compiler that uses the built-in outer-interpreter and supports literal values in TARG colon words --- not floating-point literals, anyway.

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.forth

csiph-web