Return to ckwart.doc CVS log

Up to [CSRG BSD Unix] / 43BSD / contrib / kermit

Annotation of 43BSD/contrib/kermit/ckwart.doc, revision 1.1.1.1

1.1       root        1: WART
                      2: 
                      3: Wart is a program that implements a small subset of the Unix 'lex' lexical
                      4: analyzer generator.  Unlike lex, wart may be distributed without requirement
                      5: for a Unix license.  Wart was written by Jeff Damens at the Columbia University
                      6: Center of Computing Activities to facilitate development of Unix Kermit.
                      7: 
                      8: Wart is intended for production of state table switchers.  It allows a set of
                      9: states to be defined, along with a function for getting input, and a table of
                     10: state transitions.  A C program is generated which performs actions and
                     11: switches states based on the current state and the input.
                     12: 
                     13: The following short program demonstrates some of the capabilities and
                     14: limitations of Wart.  The program accepts from the command line a binary
                     15: number, preceded by an optional minus sign, and optionally containing a
                     16: fractional part.  It prints the decimal equivalent.
                     17: 
                     18: #include <stdio.h>
                     19: 
                     20: int state, s = 1, m = 0, d;
                     21: float f;
                     22: char *b;
                     23: 
                     24: %states sign mantissa fraction             /* Declare wart states */
                     25: 
                     26: %%                                         /* Begin state table */
                     27: <sign>-      { s = -1; BEGIN mantissa; }    /* Look for sign */
                     28: <sign>0      { m = 0;  BEGIN mantissa; }    /* Got digit, start mantissa */
                     29: <sign>1      { m = 1;  BEGIN mantissa; }
                     30: <sign>.      { fatal("bad input"); }       /* Detect bad format */
                     31: <mantissa>0  { m *= 2; }                   /* Accumulate mantissa */
                     32: <mantissa>1  { m = 2 * m + 1; }
                     33: <mantissa>$  { printf("%d\n", s * m); return; }
                     34: <mantissa>.  { f = 0.0; d = 1; BEGIN fraction; }    /* Start fraction */
                     35: <fraction>0  { d *= 2; }                           /* Accumulate fraction */
                     36: <fraction>1  { d *= 2; f += 1.0 / d; }
                     37: <fraction>$  { printf("%f\n", s * (m + f) ); return; }
                     38: <fraction>.  { fatal("bad input"); }
                     39: %%
                     40: 
                     41: input() {                                  /* Define input() function */
                     42:     int x;
                     43:     return(((x = *b++) == '\0') ? '$' : x );
                     44: }
                     45: 
                     46: fatal(s) char *s; {                        /* Error exit */
                     47:     fprintf(stderr,"fatal - %s\n",s);
                     48:     exit(1);
                     49: }
                     50: 
                     51: main(argc,argv) int argc; char **argv; {    /* Main program */
                     52:     if (argc < 1) exit(1);
                     53:     b = *++argv;
                     54:     state = sign;                          /* Initialize state */
                     55:     wart();                                /* Invoke state switcher */
                     56:     exit(0);                               /* Done */
                     57: }
                     58: 
                     59: The wart program accepts as input a C program containing lines that start
                     60: with "%" or sections delimited by "%%".  The directive "%states" declares
                     61: the program's states.  The section enclosed by "%%" markers is the state
                     62: table, with entries of the form
                     63: 
                     64:   <state>X { action }
                     65: 
                     66: which is read as "if in state <state> with input X perform { action }"
                     67: 
                     68: The optional <state> field tells the current state or states the program must
                     69: be in to perform the indicated action.  If no state is specified, then it
                     70: means the action will be performed regardless of the current state.  If more
                     71: than one state is specifed, then the action will be performed in any of the
                     72: listed states.  Multiple states are separated by commas.
                     73: 
                     74: The required input field consists of a single literal character.  When in
                     75: the indicated state, if the input is the specified character, then the
                     76: associated action will be performed.  The character '.' matches any input
                     77: character.  No pattern matching or range notation is provided.  The input
                     78: character is obtained from the input() function, which you must define.  It
                     79: should be alphanumeric, or else one of the characters ".% -$@" (quotes not
                     80: included).  Note that the program above recognize the binary point '.'
                     81: through a ruse.
                     82: 
                     83: The action is a series of zero or more C language statements, enclosed in
                     84: curly braces.
                     85: 
                     86: The BEGIN macro is defined simply to be "state = ", as in lex.
                     87: 
                     88: The wart() function is generated by the wart program based on the state
                     89: declarations and the state transition table.  It loops through calls to
                     90: input(), using the result to index into a big case statement it has created
                     91: from the state table.
                     92: 
                     93: Wart is invoked as follows:
                     94: 
                     95:        wart          (Input from stdin, output to stdout)
                     96: 
                     97:        wart fn1      (Input from fn1, output to stdout)
                     98: 
                     99:        wart fn1 fn2  (Input from fn1, output to fn2.  Example:  wart a.w a.c)
                    100: 
                    101: Wart programs have the conventional filetype '.w'.

This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.