/*
** MBPRE.C  v 1.2
**
** MBasic PREprocessor
**
** 28 Jul 86	Modified to allow the use of  a semicolon, or the
**   v 1.2	word REM,  at the beginning of a line  to specify
**		that the line is a comment or remark line.  Blank
**		lines are still ignored as in the original.
**					G. F. Reding   [72436,45]
**
** 07 Nov 85	Typed from Microsystems magazine, May 1984 issue.
**   v 1.1	Modified for compilation with BDS-C v1.50.  Added
**		strip of high bit from input strings. Document or
**		nondocument mode of text editor may be used. Some
**		minor changes made, plus more comments added.
**					G. F. Reding   [72436,45]
**
** Original by	James L. Shearer, Angola, IN  46703.
**		For compilation with C80, by Software Toolworks.
**					Thanks very much! <GFR>
**
**
** This program allows the creation of a MBasic programs in stuctured
** format using a text editor in either the document or non-document mode.
** Line numbers are NOT USED.  Instead, labels mark the entry/exit points
** of routines.  All labels are 8-character strings (or shorter) whose
** first character MUST be the '@' character and subsequent characters
** may be letters, numerals, or periods, inclusive.  Labels longer than
** 8 characters are allowed but would be truncated, so only the first 8
** characters would be recognized.  Following are some examples:
**
**	Acceptable	Unacceptable
**
**	@PRINT.1	@WRITE#4	( # is an illegal char )
**	@EXIT		EXIT.6		( no leading @ )
**	@age.sr		@POSITION.35	( too long, would be truncated)
**
** Simply, write your program as you would in assembly, C-language, etc.
** Lines BEGINNING with 'REM' are remark lines, as in basic.  These such
** lines are not included in the resultant file created by this program.
**
** If a REM is included near the end of a program line, then such remark
** is retained in the new file, which could be helpful at  odd spots and
** for some areas of difficult code, etc.
**
** +++ NEW +++ Lines beginning with a semicolon, as in assembly language
** programs, are also considered as "remark lines".
**
** The capability to use the (original) REM is retained in this version,
** should anyone prefer to use it instead of the (new) semicolon, and/or
** to allow persons time to convert their old programs to the new format
** of using just a semicolon at the beginning of a remark/comment line.
**
** +++ FUTURE +++ Version 1.3 will NOT use the REM, but will be ONLY for
** remark/comment lines that start with the semicolon.  Other minor mods
** may also be included.
**
**		Usage: mbpre oldfile.ext newfile.ext <cr>
**
** NOTE:  Be sure no file by your proposed newfile.ext name exists when
** you use this as it would be written over.  Suggest as extentions for
** the oldfile (source) and newfile (destination) respectively to be
**			  .MBP   and   .BAS
**
*/

#include	<bdscio.h>			/* standard header file	*/

#define		MAXLEN 255			/* Max line length	*/
#define		MAXLBL 500			/* Room for 500 labels	*/

char		ibuf[BUFSIZ],			/* input file buffer	*/
		obuf[BUFSIZ],			/* output file buffer	*/
		label[9*MAXLBL];		/* arrays for labels	*/

int		labelno[MAXLBL],		/* arrays for line nbrs	*/
		lblcnt,				/* count of labels	*/
		lineno;				/* line number variable	*/


main(argc,argv)
int argc;					/* arg count		*/
char **argv;					/* arg vector		*/
{

	char	line[MAXLEN],			/* temp line string	*/
		word[MAXLEN];			/* temp word string	*/
	int	pass,				/* in-file read counter	*/
		len;				/* length of line read	*/

		printf("\nMBPRE  MBasic PREprocessor v1.2\n");

	if (argc != 3)				/* allow only 2 args	*/
	{
		printf("Usage: mbpre oldfile.ext newfile.ext <cr>\n");
		exit();
	}

	if (fopen(argv[1],ibuf) == ERROR)	/* try to open file	*/
	{
		printf("Can't open: %s\n",argv[1]);
		exit();
	}

	if (fcreat(argv[2],obuf) == ERROR)	/* create 2nd file	*/
	{
		printf("Can't create: %s\n",argv[2]);
		exit();
	}

/*
** On either pass, test for REM lines and drop any that are found.
** "Remark lines" may also be any line which begins with the first
** character of a semicolon in addition to the alternative of a line
** which begins with the first 3 characters of REM.
**
** On first pass, assign line numbers for the labels and build the
** label table as the labels are found.  On the second pass, write
** numbered program lines to out-file, eliminating any unnecessary
** white-space from beginning of each program line and replacing any
** labels found in the program line with its appropriate line number.
*/

	pass = 1;			/* show this is the first pass	*/

	while (pass != 3)				/* use only 2x	*/
	{
		if (pass == 2)
		{
			/* if pass 2, re-open in-file for second read	*/

			if (fopen(argv[1],ibuf) == ERROR)
			{
				printf("Can't open: %s\n",argv[1]);
				exit();
			}
		}

		lineno = 0;		/* initialize line nbr to zero	*/
		
		while ((len = getln(line,MAXLEN)) > 0)	/* do until end	*/
		{
			getwrd(word,line);

	/* this next conditional makes sure that blank lines or lines which
	   begin with either a semicolon or the letters REM are not counted
	   and consequently are not sent to the output file		*/

	if ((len > 2) && (strcmp(word,";") != 0) && (strcmp(word,"REM") != 0))
			{
				/* if not REM line, assign a line nbr,	*/
				/* build or write depending on pass nbr	*/

				lineno = lineno + 10;	/* give line #	*/

				if (pass == 1)
					{
					if (*line == '@')
					/* build table if label		*/
						bldlabel(line,lineno);
					}

				else writeln(line,lineno);
			}
		}

		fclose(ibuf);		/* close at end of each pass	*/
		pass = pass + 1;	/* show doing another pass	*/
	}

	int j;
	putc(CPMEOF,obuf);				/* add EOF char	*/
	fclose(obuf);				/* close with autoflush	*/
	for (j = 0; j <= lblcnt; ++j)			/* give summary	*/
		printf("\n%4d %10s",labelno[j],label+9*j);
	printf("\n");
	exit();

}

/*
** Subroutines used above
*/

getln(s,lim)				/* Reads line from file f1 to s	*/
int lim;				/* Returns line len (0 if EOF)	*/
char s[];
{
	int c, i, mask;
	mask = 127;				/* high order bit mask	*/
	i = 0;
	do (c=getc(ibuf));			/* pass white space	*/
		while (isspace(c));
	if (c != EOF && c != CPMEOF &&c != '\n')
	{
		/* strip hi order bit of characters by AND with mask	*/
		s[i++] = (c & mask);
		while (--lim > 0 && (c=getc(ibuf)) != EOF && c != '\n')
			s[i++] = (c & mask);
	}
	if (c == '\n')
		s[i++] = c;
	s[i] = '\0';
	return(i);
}

bldlabel(s,n)			/* Put label from line s and line	*/
char *s;			/* number n into arrays.		*/
int n;
{
	if (lblcnt <= MAXLBL);
	{
		wrdcpy(label+9*lblcnt,s);
		label[9*lblcnt+8] = '\0';	/* limit label to 8 char */
		labelno[lblcnt] = n;
		lblcnt = lblcnt + 1;
	}
}

writeln(s,n)		/* Write line s to outfile inserting line	*/
int n;			/* number n, and replace internal labels	*/
char s[];		/* with appropriate line number.		*/
{
	int i, j;
	char t[MAXLEN];
	i = 0;
	if (s[i] == '@')			/* label if 1st char @	*/
	{
		while (!isspace(s[i++]))	/* and we pass it	*/
			;
		while (isspace(s[i]))		/* and white space	*/
			++i;
	}
	fprintf(obuf,"%d ",n);			/* write line number	*/
	while (s[i] != '\n')
	{
		if (s[i] == '"')		/* check for quotes	*/
		{
			putc(s[i],obuf);	/* write first " and	*/
			while (s[++i] != '"')	/* and all in between	*/
				putc(s[i],obuf);
			putc(s[i++],obuf);	/* write last "		*/
		}
		else if (s[i] == '@')		/* check for label	*/
		{
			getwrd(t,&s[i]);	/* get label into t	*/
			i = i + strlen(t);	/* move ptr past label	*/
			t[8] = '\0';		/* only 8 char signif	*/

			for (j = 0; j <= lblcnt; j++)
				if (strcmp(t,label+9*j) == 0)
				{	/* write line # if match */
					fprintf(obuf,"%d",labelno[j]);
					break;
				}

			if (j > lblcnt )	/* no match		*/
			{
				printf("\n##### No match for label ");
				printf("%s on line %d\n",t,n);
				fprintf(obuf,"?%s?",t);
			}
		}
		else
			putc(s[i++],obuf);	/* regular character	*/
	}
	putc(s[i],obuf);			/* end-of-line char(s)	*/
}

getwrd(t,s)			/* Get next word from s, put it in t.	*/
char s[],t[];			/* For our purposes, words start with @	*/
{				/* or alpha and then alpha, digits,	*/
	int i, j;		/* embedded '.' and optional final char	*/
	i = 0;

	if ((s[i] == ';') && (i == 0))		/* if start is a semi	*/
		i++;				/* then get only it	*/

	else if (s[i] == '@' || isalpha(s[i]))	/* otherwise handle as	*/
						/* normal line and get	*/
		{				/* first word of line	*/
		++i;
		while (isalpha(s[i]) || isdigit(s[i]) || s[i] == '.')
				i++;
		if (s[i]=='%' || s[i]=='!' || s[i]=='#' || s[i]=='$')
				i++;
		}
	for (j = 0; j <= i; j++)
		t[j] = s[j];
	t[i] = '\0';
}

wrdcpy(s,t)		/* copy next word (see getwrd) from t to s	*/
char *s, *t;
{
	while (isalpha(*t) || isdigit(*t) || *t == '@' || *t == '.')
	{
		*s = *t;
		s++;
		t++;
	}
}