intro to scanning

[Gene Gaunt <issgg@xxxxxxx>]

I write MI in lowercase.  My scanner capitalizes where appropriate, 
skipping literals and comments.  It converts C++ style comments to CL-style 
comments.  Here's a before-and-after example.

cmpbla(b)  DayOfWeek, "Mc'Donald's" / eq(ReadAgain);  -- jump if found
CMPBLA(B)  DAYOFWEEK, "Mc'Donald's" / EQ(READAGAIN);  /* jump if found */

Here's a snip of scanner code.  Put it in your QPRCRTPG front end.  Point 
?SCAN to your MI source string.  Calculate LENGTH, and let it run!  The 
first state is 1.  When the scanner enters state 0, it capitalizes the 
current token.  (If you want C++ style comments, add another table row for 
"dash"; add two more table columns for "the dash state" and "the dash-dash 
state"; and add executable code to handle "entering the dash-dash state" 
event.  Overlay the dash-dash with slash-asterisk, and append 
asterisk-slash to the current source line.)

Next, think how you'd write a general-purpose program that accepts language 
definitions and outputs program source like this.  Then you're on your way 
to creating little languages for AS/400, for example, a menu-bar pull-down 
window language, a panel group driver language, an infix-to-postfix 
converter front end for QQQQUERY.  That's stuff people will pay big money 
for!

DCL MSPPTR ?SCAN;
DCL DD     SCANBYTE   BAS(?SCAN) CHAR(1);
DCL DD     LENGTH     AUTO BIN(4);
DCL DD     INPUT      AUTO BIN(2) INIT(0);
DCL DD     INPUT2     DEF(INPUT) POS(2) CHAR(1);
DCL DD     STATE      AUTO BIN(2) INIT(1);
DCL DD     NEXT(0:63) AUTO INIT( 1,  1,  2,  1,  4,  4,  6,  7,    /* other 
*/
                                 2,  2,  2,  1,  4,  4,  6,  7,    /*   < 
  */
                                 1,  1,  2,  4,  5,  4,  6,  7,    /*   * 
  */
                                 3,  3,  2,  1,  4,  1,  6,  7,    /*   / 
  */
                                 1,  1,  1,  1,  4,  4,  6,  7,    /*   > 
  */
                                 6,  6,  2,  1,  4,  4,  1,  7,    /*   ' 
  */
                                 7,  7,  2,  1,  4,  4,  6,  1,    /*   " 
  */
                                 0,  0,  2,  0,  4,  4,  6,  7);   /* lower 
*/

DCL DD GROUPS AUTO                   CHAR(256) INIT((256) X"01");  /* other 
*/
DCL DD *      DEF(GROUPS) POS(H"4D") CHAR( 1 ) INIT(( 1 ) X"02");  /*   < 
  */
DCL DD *      DEF(GROUPS) POS(H"5D") CHAR( 1 ) INIT(( 1 ) X"03");  /*   * 
  */
DCL DD *      DEF(GROUPS) POS(H"62") CHAR( 1 ) INIT(( 1 ) X"04");  /*   / 
  */
DCL DD *      DEF(GROUPS) POS(H"6F") CHAR( 1 ) INIT(( 1 ) X"05");  /*   > 
  */
DCL DD *      DEF(GROUPS) POS(H"7E") CHAR( 1 ) INIT(( 1 ) X"06");  /*   ' 
  */
DCL DD *      DEF(GROUPS) POS(H"80") CHAR( 1 ) INIT(( 1 ) X"07");  /*   " 
  */
DCL DD *      DEF(GROUPS) POS(H"82") CHAR( 9 ) INIT(( 9 ) X"08");  /*  a-i 
 */
DCL DD *      DEF(GROUPS) POS(H"92") CHAR( 9 ) INIT(( 9 ) X"08");  /*  j-r 
 */
DCL DD *      DEF(GROUPS) POS(H"A3") CHAR( 8 ) INIT(( 8 ) X"08");  /*  s-z 
 */

LOOP: XLATEWT   INPUT2, SCANBYTE, GROUPS;
      CAI       INPUT, STATE, INPUT, 8;
      CPYNV(B)  STATE, NEXT(INPUT) / NZER(SKIP);
      OR(S)     SCANBYTE, X"40";
SKIP: ADDSPP    ?SCAN, ?SCAN, 1;
      SUBN(SB)  LENGTH, 1 / POS(LOOP);

+---
| This is the MI Programmers Mailing List!
| To submit a new message, send your mail to MI400@midrange.com.
| To subscribe to this list send email to MI400-SUB@midrange.com.
| To unsubscribe from this list send email to MI400-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: dr2@cssas400.com
+---