M4BASIC

[alexfish]

Hi Peter

Have been doing some testing on M4BASIC

have noted that RETURN has +0 , this throws an err if trying to pass a widget pointer

on the file open , can not allocate more than one file , also can not re-open the file


BR
Alex

[Pjot]

Thanks Alex,

I will take a look at it.

BR
Peter

[alexfish]

Hi Peter

been playing around with CONCAT

looks like the m4 version does not like certain chars

in short did some bits in bacon

the first line a$ is from part of the exec script

the code will output some bits that can be tested with M4BASIC

also can adapt the first a $ for M4BASIC, then will give approximation of some limitations.

not sure what actual problems is since read and concat of file appears to be working see second code block

BR
Alex


a$=CONCAT$("tail -n +${LINE%%:*} $0 | ${AWK} '{if ($0 != ",CHR$(34),"BASIC",CHR$(34),"){split($0, byte); for(i=1; i <= NF; i++) printf(",CHR$(34),"%c",CHR$(34)," byte[i]);} }' > ${TEMPDIR}/basic.m4.gz")

a$=CONCAT$("tail -n +${LINE%%:*} $0 | ${AWK} '{if ($0 != ",CHR$(34),"BASIC",CHR$(34),"){split($0, byte);")
FOR t= 1 TO LEN(a$)
new$=CONCAT$(new$,"CHR(",STR$(ASC(MID$(a$,t,1))),"),")
NEXT
PRINT new$
PRINT "\n--------------------------------------\n"
a$=CONCAT$("for(i=1; i <=; i++) printf(",CHR$(34),"%c",CHR$(34)," byte[i]);} }' > ${TEMPDIR}/basic.m4.gz")
FOR t= 1 TO LEN(a$)
new$=CONCAT$(new$,"CHR(",STR$(ASC(MID$(a$,t,1))),"),")
NEXT
PRINT new$
PRINT "\n--------------------------------------\n"
a$=CONCAT$("(i=1; i <=; i++) printf(",CHR$(34),"%c",CHR$(34)," byte[i]);} }' > ${TEMPDIR}/basic.m4.gz")
FOR t= 1 TO LEN(a$)
new$=CONCAT$(new$,"CHR$(",STR$(ASC(MID$(a$,t,1))),"),")
NEXT
PRINT new$
new$=CONCAT$(CHR$(105),CHR$(61),CHR$(49),CHR$(59),CHR$(32),CHR$(105),CHR$(32),CHR$(60),CHR$(61),CHR$(59),CHR$(32),CHR$(105),CHR$(43),CHR$(43),CHR$(41),CHR$(32),CHR$(112),CHR$(114),CHR$(105),CHR$(110),CHR$(116),CHR$(102),CHR$(40),CHR$(34),CHR$(37),CHR$(99),CHR$(34),CHR$(32),CHR$(98),CHR$(121),CHR$(116),CHR$(101),CHR$(91),CHR$(105),CHR$(93),CHR$(41),CHR$(59),CHR$(125),CHR$(32),CHR$(125),CHR$(39),CHR$(32),CHR$(62),CHR$(32),CHR$(36),CHR$(123),CHR$(84),CHR$(69),CHR$(77),CHR$(80),CHR$(68),CHR$(73),CHR$(82),CHR$(125),CHR$(47),CHR$(98),CHR$(97),CHR$(115),CHR$(105),CHR$(99),CHR$(46),CHR$(109),CHR$(52),CHR$(46),CHR$(103),CHR$(122))
PRINT new$

new$=CONCAT$(CHR$(105),CHR$(61),CHR$(49),CHR$(59),CHR$(32),CHR$(105),CHR$(32),CHR$(60),CHR$(61),CHR$(59),CHR$(32),CHR$(105),CHR$(43),CHR$(43),CHR$(41),CHR$(32),CHR$(112),CHR$(114),CHR$(105),CHR$(110),CHR$(116),CHR$(102),CHR$(40),CHR$(34),CHR$(37),CHR$(99),CHR$(34),CHR$(32),CHR$(98),CHR$(121),CHR$(116),CHR$(101),CHR$(91),CHR$(105),CHR$(93),CHR$(41),CHR$(59),CHR$(125),CHR$(32),CHR$(125),CHR$(39),CHR$(32),CHR$(62),CHR$(32),CHR$(36),CHR$(123),CHR$(84),CHR$(69),CHR$(77),CHR$(80),CHR$(68),CHR$(73),CHR$(82),CHR$(125),CHR$(47),CHR$(98),CHR$(97),CHR$(115),CHR$(105),CHR$(99),CHR$(46),CHR$(109),CHR$(52),CHR$(46),CHR$(103),CHR$(122))
PRINT new$



the read in of all the file

DECLARE name TYPE char*
DECLARE concat_name TYPE char*
DECLARE myfile TYPE FILEPTR
DECLARE myfile2 TYPE FILEPTR
DECLARE isp TYPE long
DECLARE ln TYPE long



name = "Hello cruel world";


PRINT (LEN(name)) FORMAT "%ld\n"
REM init the variable concat_name to "" else get garbage at start of concat
LET concat_name=""

PRINT "\n File read result \n --------------- \n"
IF FILEEXISTS("bm4.sh") THEN

OPEN "bm4.sh" FOR READING AS myfile

WHILE NOT(ENDFILE(myfile)) DO
READLN name FROM myfile

LET isp = INSTR(name, "=")
IF isp >0 THEN
PRINT "= found\n"
ENDIF

LET concat_name=CONCAT(concat_name,name)
WEND

CLOSE myfile

PRINT concat_name NL

ENDIF


PRINT " *-----ppp---------*\n -- * DONE *--\n"

[Pjot]

Hi Alex,

For your last program, please note that M4 does not do any parsing. So it cannot see if a keyword is mentioned in a string between double quotes, or not.

For example, these lines will go wrong:

FOR t= 1 TO LEN(a$)
  new$=CONCAT$(new$,"CHR(",STR$(ASC(MID$(a$,t,1))),"),")
NEXT

Here, CHR is mentioned within double quotes. However, M4 cannot see this, and will think it is a macro, expanding it right away, which gives unpredictable results.

BR
Peter

[alexfish]

Hi Peter

think post not easy to understand

Run the code under bacon , all bits work

use the bacon bacongui shell , you will notice the bits that can be copied into the
M4BASIC CONCAT

I am aware the $ is not required in M4BASIC

adapt the first line of the concat code in the post and test in M4BASIC

also test the the CHR bits from the output of bacon shell they

once code is in M4BASIC of concating the chr(*) bits
try to print it

then possible will understand

1. yes i am aware of the parser limitations of m4 , also , hint hint lex yak and bison
2. think also there be a problem with concat method or the print.

BR
Alex

[Pjot]

Alex,

For any assignment you must use LET. In case of strings the LET statement just assigns a pointer. So then it is an assignment 'by reference'.

Therefore, if you want to assign a string to a variable 'by value', then you need to use SET instead. This is the default BASIC way of assigning strings to variables. So, all your CONCAT assignments must be preceded with SET first.

Thanks for looking into the M4Basic project. It is an experiment to see how far we can get with plain M4. Interestingly, only a few lines of code are needed to already get a very powerful converter, be it not a portable one; it only works with GNU C and the Boehm Grabage collector.

Of course I know Lex and Yacc (or Flex and Bison, their GNU counterparts). Especially an implementation based on Flex would be interesting!

BR
Peter

[alexfish]

Hi Peter

Found the limitation of M4BASIC CONCAT

Yes Understand limitation of certain chars , IE in Bacon = ", this we know , one uses CHR$

M4 BASIC has a few more , so one has to experiment to find those bits,

The main limitation on the concat iterator

Looks like the MAX ITER = 85

as a workaround one can concat up to 85 bit then can add to it

IE This works

SET my_concat = CONCAT(CHR(116),CHR(97),CHR(105),CHR(108),CHR(32),CHR(45),CHR(110),CHR(32),CHR(43),CHR(36),CHR(123),CHR(76),CHR(73),CHR(78),CHR(69),CHR(37),CHR(37),CHR(58),CHR(42),CHR(125),CHR(32),CHR(36),CHR(48),CHR(32),CHR(124),CHR(32),CHR(36),CHR(123),CHR(65),CHR(87),CHR(75),CHR(125),CHR(32),CHR(39),CHR(123),CHR(105),CHR(102),CHR(32),CHR(40),CHR(36),CHR(48),CHR(32),CHR(33),CHR(61),CHR(32),CHR(34),CHR(66),CHR(65),CHR(83),CHR(73),CHR(67),CHR(34),CHR(41),CHR(123),CHR(115),CHR(112),CHR(108),CHR(105),CHR(116),CHR(40),CHR(36),CHR(48),CHR(44),CHR(32),CHR(98),CHR(121),CHR(116),CHR(101),CHR(41),CHR(59),CHR(32),CHR(102),CHR(111),CHR(114),CHR(40),CHR(105),CHR(61),CHR(49),CHR(59),CHR(32),CHR(105),CHR(32),CHR(60),CHR(61),CHR(32))

SET my_concat = CONCAT(my_concat,CHR(78),CHR(70),CHR(59),CHR(32),CHR(105),CHR(43),CHR(43),CHR(41),CHR(32),CHR(112),CHR(114),CHR(105),CHR(110),CHR(116),CHR(102),CHR(40),CHR(34),CHR(37),CHR(99),CHR(34),CHR(32),CHR(98),CHR(121),CHR(116),CHR(101),CHR(91),CHR(105),CHR(93),CHR(41),CHR(59),CHR(125),CHR(32),CHR(125),CHR(39),CHR(32),CHR(62),CHR(32),CHR(36),CHR(123),CHR(84),CHR(69),CHR(77),CHR(80),CHR(68),CHR(73),CHR(82),CHR(125),CHR(47),CHR(98),CHR(97),CHR(115),CHR(105),CHR(99),CHR(46),CHR(109),CHR(52),CHR(46),CHR(103),CHR(122))

PRINT my_concat


as a side note LET statement also works.

In regards to lex , have done some work on that side as regards code completion for the likes of Elementary widgets + the lower level libs
IE if statement = WINDOW , then result = rest of lines to complete the bits .

of EFL ,
EFL has a lot of low end libs that can fill the gap between where BACON is now and what M4 basic lacs at Present.

if one wants to use Elementary Widgets or just the Evas Canvas those libs are there at default.

M4BASIC + EFL looks like a good partnership .

End Result , very lightweight execs.

BR
ALEX

[Pjot]

Hi Alex,

Interesting that GNU M4 cannot handle more than 85 arguments in a macro! Obviously, this is an M4 limitation.

Note that the original M4 cannot handle more than 9 arguments.

BR
Peter

[alexfish]

Hi Peter

those bits had me foxed , but at not sure to the actual limitation , is a stack size problem? may be worth looking up to see if it is.

anyway, thought would do a bit of daft Flex code , had already prepped this so posting for any user interested in Flex.

Very basic . so posting the whole antic as is , just kind of thinking M4BASIC code make light work of coding bits for Flex.

Have Fun +
BR
Alex
-------------------------------
Hi Peter and All

thought show why use 'Flex'

for me one can structure the CONDITION code just like c ;

here have chopped up some so can see what going on as regards adding sub routines
the condition code , very basic and none nonsensical , but there are not may examples of
how to do it.

much same as m4 not that I want to start learning it either ! , but I find this easier than M4 hieroglyphs.

some code , compile commands then the code :: the a.out < points to the file to parse


flex myexample.l
gcc lex.yy.c -lfl -o myexample
./myexample ./a.out < testfile.bac


some ridiculus code..

and examples test file
%{
#include <stdio.h>
#include <string.h>
#include "y.tab.h"
#include <math.h>


%}


int num_lines = 0, num_chars = 0 , is_win=0 ;
char* on = "ON";



%%

[a-z]  printf( "%s", yytext )  ;

" " printf( "%s", yytext )  ;

"WINDOW" { printf( "%s%s", yytext," ("  ) ;
 printf( "%s", " `" ) ;
is_win=1;

void Do_Something(){
puts (" do something here::  Bye");
}
if (is_win)
{
puts ("Parsing Window blagh blagh");
}
Do_Something();
}

[0123456789]+ {printf( "%d",atoi( yytext ) )
;}

\n puts ("");

%%

main(){

 yylex();

}


a bit to test the out put

testfile.bac

win = WINDOW("window", 600 , 600)

chopping the bits


Bye For now