2 // RMAC - Reboot's Macro Assembler for the Atari Jaguar Console System
3 // TOKEN.C - Token Handling
4 // Copyright (C) 199x Landon Dyer, 2011-2012 Reboot and Friends
5 // RMAC derived from MADMAC v1.07 Written by Landon Dyer, 1986
6 // Source Utilised with the Kind Permission of Landon Dyer
15 #define DECL_KW // Declare keyword arrays
16 #define DEF_KW // Declare keyword values
17 #include "kwtab.h" // Incl generated keyword tables & defs
20 int lnsave; // 1; strcpy() text of current line
21 int curlineno; // Current line number
22 int totlines; // Total # of lines
23 int mjump_align = 0; // mjump alignment flag
24 char lntag; // Line tag
25 char * curfname; // Current filename
26 char tolowertab[128]; // Uppercase ==> lowercase
27 char hextab[128]; // Table of hex values
28 char dotxtab[128]; // Table for ".b", ".s", etc.
29 char irbuf[LNSIZ]; // Text for .rept block line
30 char lnbuf[LNSIZ]; // Text of current line
31 WORD filecount; // Unique file number counter
32 WORD cfileno; // Current file number
33 TOKEN * tok; // Ptr to current token
34 TOKEN * etok; // Ptr past last token in tokbuf[]
35 TOKEN tokeol[1] = {EOL}; // Bailout end-of-line token
36 char * string[TOKBUFSIZE*2]; // Token buffer string pointer storage
38 // File record, used to maintain a list of every include file ever visited
39 #define FILEREC struct _filerec
49 INOBJ * cur_inobj; // Ptr current input obj (IFILE/IMACRO)
50 static INOBJ * f_inobj; // Ptr list of free INOBJs
51 static IFILE * f_ifile; // Ptr list of free IFILEs
52 static IMACRO * f_imacro; // Ptr list of free IMACROs
54 static TOKEN tokbuf[TOKBUFSIZE]; // Token buffer (stack-like, all files)
57 ILLEG, ILLEG, ILLEG, ILLEG, // NUL SOH STX ETX
58 ILLEG, ILLEG, ILLEG, ILLEG, // EOT ENQ ACK BEL
59 ILLEG, WHITE, ILLEG, ILLEG, // BS HT LF VT
60 WHITE, ILLEG, ILLEG, ILLEG, // FF CR SO SI
62 ILLEG, ILLEG, ILLEG, ILLEG, // DLE DC1 DC2 DC3
63 ILLEG, ILLEG, ILLEG, ILLEG, // DC4 NAK SYN ETB
64 ILLEG, ILLEG, ILLEG, ILLEG, // CAN EM SUB ESC
65 ILLEG, ILLEG, ILLEG, ILLEG, // FS GS RS US
67 WHITE, MULTX, MULTX, SELF, // SP ! " #
68 MULTX+CTSYM, MULTX, SELF, MULTX, // $ % & '
69 SELF, SELF, SELF, SELF, // ( ) * +
70 SELF, SELF, STSYM, SELF, // , - . /
72 DIGIT+HDIGIT+CTSYM, DIGIT+HDIGIT+CTSYM, // 0 1
73 DIGIT+HDIGIT+CTSYM, DIGIT+HDIGIT+CTSYM, // 2 3
74 DIGIT+HDIGIT+CTSYM, DIGIT+HDIGIT+CTSYM, // 4 5
75 DIGIT+HDIGIT+CTSYM, DIGIT+HDIGIT+CTSYM, // 6 7
76 DIGIT+HDIGIT+CTSYM, DIGIT+HDIGIT+CTSYM, // 8 9
78 MULTX, MULTX, MULTX, STSYM+CTSYM, // < = > ?
80 MULTX, STSYM+CTSYM+HDIGIT, // @ A
81 (char)((BYTE)DOT)+STSYM+CTSYM+HDIGIT, STSYM+CTSYM+HDIGIT, // B C
82 STSYM+CTSYM+HDIGIT, STSYM+CTSYM+HDIGIT, // D E
83 STSYM+CTSYM+HDIGIT, STSYM+CTSYM, // F G
84 STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, // H I J K
85 (char)((BYTE)DOT)+STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, // L M N O
87 STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, (char)((BYTE)DOT)+STSYM+CTSYM, // P Q R S
88 STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, (char)((BYTE)DOT)+STSYM+CTSYM, // T U V W
89 STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, SELF, // X Y Z [
90 SELF, SELF, MULTX, STSYM+CTSYM, // \ ] ^ _
92 ILLEG, STSYM+CTSYM+HDIGIT, // ` a
93 (char)((BYTE)DOT)+STSYM+CTSYM+HDIGIT, STSYM+CTSYM+HDIGIT, // b c
94 STSYM+CTSYM+HDIGIT, STSYM+CTSYM+HDIGIT, // d e
95 STSYM+CTSYM+HDIGIT, STSYM+CTSYM, // f g
96 STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, // h i j k
97 (char)((BYTE)DOT)+STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, // l m n o
99 STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, (char)((BYTE)DOT)+STSYM+CTSYM, // p q r s
100 STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, (char)((BYTE)DOT)+STSYM+CTSYM, // t u v w
101 STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, SELF, // x y z {
102 SELF, SELF, SELF, ILLEG // | } ~ DEL
105 // Names of registers
106 static char * regname[] = {
107 "d0", "d1", "d2", "d3", "d4", "d5", "d6", "d7",
108 "a0", "a1", "a2", "a3", "a4", "a5", "a6", "a7",
109 "pc", "ssp", "usp", "sr", "ccr"
112 static char * riscregname[] = {
113 "r0", "r1", "r2", "r3", "r4", "r5", "r6", "r7",
114 "r8", "r9", "r10", "r11", "r12", "r13", "r14", "r15",
115 "r16", "r17", "r18", "r19", "r20", "r21", "r22", "r23",
116 "r24", "r25", "r26", "r27", "r28", "r29", "r30", "r31"
121 // Initialize Tokenizer
123 void InitTokenizer(void)
126 char * htab = "0123456789abcdefABCDEF"; // Hex character table
128 lnsave = 0; // Don't save lines
129 curfname = ""; // No file, empty filename
130 filecount = (WORD)-1;
131 cfileno = (WORD)-1; // cfileno gets bumped to 0
143 // Initialize hex, "dot" and tolower tables
148 tolowertab[i] = (char)i;
151 for(i=0; htab[i]!=EOS; i++)
152 hextab[htab[i]] = (char)((i < 16) ? i : i - 6);
154 for(i='A'; i<='Z'; i++)
155 tolowertab[i] |= 0x20;
157 // These characters are legal immediately after a period
158 dotxtab['b'] = DOTB; // .b .B .s .S
162 dotxtab['w'] = DOTW; // .w .W
164 dotxtab['l'] = DOTL; // .l .L
166 dotxtab['i'] = DOTI; // .i .I (???)
171 void SetFilenameForErrorReporting(void)
175 // Check for absolute top filename (this should never happen)
178 curfname = "(*top*)";
182 FILEREC * fr = filerec;
184 // Advance to the correct record...
185 while (fr != NULL && fnum != 0)
191 // Check for file # record not found (this should never happen either)
194 curfname = "(*NOT FOUND*)";
198 curfname = fr->frec_name;
203 // Allocate an IFILE or IMACRO
205 INOBJ * a_inobj(int typ)
211 // Allocate and initialize INOBJ first
213 inobj = malloc(sizeof(INOBJ));
217 f_inobj = f_inobj->in_link;
222 case SRC_IFILE: // Alloc and init an IFILE
224 ifile = malloc(sizeof(IFILE));
228 f_ifile = f_ifile->if_link;
231 inobj->inobj.ifile = ifile;
233 case SRC_IMACRO: // Alloc and init an IMACRO
234 if (f_imacro == NULL)
235 imacro = malloc(sizeof(IMACRO));
239 f_imacro = f_imacro->im_link;
242 inobj->inobj.imacro = imacro;
244 case SRC_IREPT: // Alloc and init an IREPT
245 inobj->inobj.irept = malloc(sizeof(IREPT));
246 DEBUG printf("alloc IREPT\n");
250 // Install INOBJ on top of input stack
251 inobj->in_ifent = ifent; // Record .if context on entry
252 inobj->in_type = (WORD)typ;
253 inobj->in_otok = tok;
254 inobj->in_etok = etok;
255 inobj->in_link = cur_inobj;
263 // Perform macro substitution from 'orig' to 'dest'. Return OK or some error.
264 // A macro reference is in one of two forms:
265 // \name <non-name-character>
267 // A doubled backslash (\\) is compressed to a single backslash (\).
268 // Argument definitions have been pre-tokenized, so we have to turn them back
269 // into text. This means that numbers, in particular, become hex, regardless of
270 // their representation when the macro was invoked. This is a hack.
271 // A label may appear at the beginning of the line:
272 // :<name><whitespace>
273 // (the colon must be in the first column). These labels are stripped before
274 // macro expansion takes place.
276 int ExpandMacro(char * src, char * dest, int destsiz)
279 int questmark; // \? for testing argument existence
280 char mname[128]; // Assume max size of a formal arg name
281 char numbuf[20]; // Buffer for text of CONSTs
284 char ** symbolString;
286 DEBUG { printf("ExM: src=\"%s\"\n", src); }
288 IMACRO * imacro = cur_inobj->inobj.imacro;
289 int macnum = (int)(imacro->im_macro->sattr);
292 char * dst = dest; // Next dest slot
293 char * edst = dest + destsiz - 1; // End + 1(?) of dest buffer
295 // Check for (and skip over) any "label" on the line
301 while (*s != EOS && !(chrtab[*s] & WHITE))
305 s++; // Skip first whitespace
308 // Expand the rest of the line
311 // Copy single character
319 // Do macro expansion
327 case '\\': // \\, \ (collapse to single backslash)
333 case '?': // \? <macro> set `questmark' flag
337 case '#': // \#, number of arguments
338 sprintf(numbuf, "%d", (int)imacro->im_nargs);
340 case '!': // \! size suffix supplied on invocation
341 switch ((int)imacro->im_siz)
343 case SIZN: d = ""; break;
344 case SIZB: d = ".b"; break;
345 case SIZW: d = ".w"; break;
346 case SIZL: d = ".l"; break;
350 case '~': // ==> unique label string Mnnnn...
351 sprintf(numbuf, "M%u", curuniq);
367 return error("missing argument name");
370 // \n ==> argument number 'n', 0..9
371 if (chrtab[*s] & DIGIT)
381 // Get argument name: \name, \{name}
391 while (chrtab[*s] & CTSYM);
396 for(++s; *s != EOS && *s != '}';)
400 return error("missing '}'");
407 // Lookup the argument and copy its (string) value into the
408 // destination string
409 DEBUG printf("argument='%s'\n", mname);
411 if ((arg = lookup(mname, MACARG, macnum)) == NULL)
412 return errors("undefined argument: '%s'", mname);
415 // Convert a string of tokens (terminated with EOL) back into
416 // text. If an argument is out of range (not specified in the
417 // macro invocation) then it is ignored.
418 i = (int)arg->svalue;
420 DEBUG printf("~argnumber=%d (argBase=%u)\n", i, imacro->argBase);
423 if (i < imacro->im_nargs)
428 tk = argPtrs[imacro->argBase + i];
430 tk = imacro->argument[i].token;
431 symbolString = imacro->argument[i].string;
434 // printf("ExM: Preparing to parse argument #%u...\n", i);
441 // 0 if the argument is empty or non-existant,
442 // 1 if the argument is not empty
445 if (tk == NULL || *tk == EOL)
451 *dst++ = (char)(questmark + '0');
455 // Argument # is in range, so expand it
460 // Reverse-translation from a token number to a string.
461 // This is a hack. It might be better table-driven.
464 if ((*tk >= KW_D0) && !rdsp && !rgpu)
466 d = regname[(int)*tk++ - KW_D0];
469 else if ((*tk >= KW_R0) && (*tk <= KW_R31))
471 d = riscregname[(int)*tk++ - KW_R0];
480 // d = (char *)*tk++;
483 // This fix should be done for strings too
484 d = symbolString[*tk++];
485 DEBUG printf("ExM: SYMBOL=\"%s\"", d);
490 // d = (char *)*tk++;
493 d = symbolString[*tk++];
514 // Shamus: Changing the format specifier from %lx to %ux caused
515 // the assembler to choke on legitimate code... Need to investigate
516 // this further before changing anything else here!
518 sprintf(numbuf, "$%lx", (LONG)*tk++);
579 *dst++ = (char)*(tk - 1);
584 // If 'd' != NULL, copy string to destination
588 DEBUG printf("d='%s'\n", d);
605 DEBUG { printf("ExM: dst=\"%s\"\n", dest); }
610 DEBUG printf("*** OVERFLOW LINE ***\n%s\n", dest);
611 return fatal("line too long as a result of macro expansion");
616 // Get Next Line of Text from a Macro
618 char * GetNextMacroLine(void)
620 unsigned source_addr;
622 IMACRO * imacro = cur_inobj->inobj.imacro;
623 // LONG * strp = imacro->im_nextln;
624 struct LineList * strp = imacro->im_nextln;
626 if (strp == NULL) // End-of-macro
629 // imacro->im_nextln = (LONG *)*strp;
630 imacro->im_nextln = strp->next;
631 // ExpandMacro((char *)(strp + 1), imacro->im_lnbuf, LNSIZ);
632 ExpandMacro(strp->line, imacro->im_lnbuf, LNSIZ);
634 return imacro->im_lnbuf;
639 // Get Next Line of Text from a Repeat Block
641 char * GetNextRepeatLine(void)
644 IREPT * irept = cur_inobj->inobj.irept;
645 LONG * strp = irept->ir_nextln; // initial null
647 // Do repeat at end of .rept block's string list
650 DEBUG printf("back-to-top-of-repeat-block count=%d\n", (int)irept->ir_count);
651 irept->ir_nextln = irept->ir_firstln; // copy first line
653 if (irept->ir_count-- == 0)
655 DEBUG printf("end-repeat-block\n");
659 strp = irept->ir_nextln;
662 strcpy(irbuf, (char *)(irept->ir_nextln + 1));
663 DEBUG printf("repeat line='%s'\n", irbuf);
664 irept->ir_nextln = (LONG *)*strp;
671 // Include a Source File used at the Root, and for ".include" Files
673 int include(int handle, char * fname)
681 printf("[include: %s, cfileno=%u]\n", fname, cfileno);
683 // Alloc and initialize include-descriptors
684 inobj = a_inobj(SRC_IFILE);
685 ifile = inobj->inobj.ifile;
687 ifile->ifhandle = handle; // Setup file handle
688 ifile->ifind = ifile->ifcnt = 0; // Setup buffer indices
689 ifile->ifoldlineno = curlineno; // Save old line number
690 ifile->ifoldfname = curfname; // Save old filename
691 ifile->ifno = cfileno; // Save old file number
693 // cfileno = filecount++; // Compute new file number
694 // NB: This *must* be preincrement, we're adding one to the filecount here!
695 cfileno = ++filecount; // Compute NEW file number
696 curfname = strdup(fname); // Set current filename (alloc storage)
697 curlineno = 0; // Start on line zero
699 // Add another file to the file-record
700 fr = (FILEREC *)malloc(sizeof(FILEREC));
701 fr->frec_next = NULL;
702 fr->frec_name = curfname;
705 filerec = fr; // Add first filerec
707 last_fr->frec_next = fr; // Append to list of filerecs
710 DEBUG printf("[include: curfname: %s, cfileno=%u]\n", curfname, cfileno);
717 // Pop the Current Input Level
724 INOBJ * inobj = cur_inobj;
728 // Pop IFENT levels until we reach the conditional assembly context we
729 // were at when the input object was entered.
730 while (ifent != inobj->in_ifent)
733 tok = inobj->in_otok; // Restore tok and otok
734 etok = inobj->in_etok;
736 switch (inobj->in_type)
738 case SRC_IFILE: // Pop and release an IFILE
740 printf("[Leaving: %s]\n", curfname);
742 ifile = inobj->inobj.ifile;
743 ifile->if_link = f_ifile;
745 close(ifile->ifhandle); // Close source file
746 if (verb_flag) printf("[fpop (pre): curfname=%s]\n", curfname);
747 curfname = ifile->ifoldfname; // Set current filename
748 if (verb_flag) printf("[fpop (post): curfname=%s]\n", curfname);
749 if (verb_flag) printf("[fpop: (pre) cfileno=%d ifile->ifno=%d]\n", (int)cfileno, (int)ifile->ifno);
750 curlineno = ifile->ifoldlineno; // Set current line#
751 DEBUG printf("cfileno=%d ifile->ifno=%d\n", (int)cfileno, (int)ifile->ifno);
752 cfileno = ifile->ifno; // Restore current file number
753 if (verb_flag) printf("[fpop: (post) cfileno=%d ifile->ifno=%d]\n", (int)cfileno, (int)ifile->ifno);
755 case SRC_IMACRO: // Pop and release an IMACRO
756 imacro = inobj->inobj.imacro;
757 imacro->im_link = f_imacro;
760 case SRC_IREPT: // Pop and release an IREPT
761 DEBUG printf("dealloc IREPT\n");
762 p = inobj->inobj.irept->ir_firstln;
773 cur_inobj = inobj->in_link;
774 inobj->in_link = f_inobj;
783 // Get line from file into buf, return NULL on EOF or ptr to the start of a
786 char * GetNextLine(void)
790 int readamt = -1; // 0 if last read() yeilded 0 bytes
791 IFILE * fl = cur_inobj->inobj.ifile;
795 // Scan for next end-of-line; handle stupid text formats by treating
796 // \r\n the same as \n. (lone '\r' at end of buffer means we have to
801 d = &fl->ifbuf[fl->ifind];
803 for(p=d; i<j; i++, p++)
805 d = &fl->ifbuf[fl->ifind];
807 for(p=d, i=0, j=fl->ifcnt; i<j; i++, p++)
810 if (*p == '\r' || *p == '\n')
817 break; // Need to read more, then look for '\n' to eat
818 else if (p[1] == '\n')
822 // Cover up the newline with end-of-string sentinel
831 // Handle hanging lines by ignoring them (Input file is exhausted, no
832 // \r or \n on last line)
833 // Shamus: This is retarded. Never ignore any input!
834 if (!readamt && fl->ifcnt)
841 // Really should check to see if we're at the end of the buffer! :-P
842 fl->ifbuf[fl->ifind + fl->ifcnt] = '\0';
844 return &fl->ifbuf[fl->ifind];
848 // Truncate and return absurdly long lines.
849 if (fl->ifcnt >= QUANTUM)
851 fl->ifbuf[fl->ifind + fl->ifcnt - 1] = '\0';
853 return &fl->ifbuf[fl->ifind];
856 // Relocate what's left of a line to the beginning of the buffer, and
857 // read some more of the file in; return NULL if the buffer's empty and
861 p = &fl->ifbuf[fl->ifind];
862 d = &fl->ifbuf[fl->ifcnt & 1];
864 for(i=0; i<fl->ifcnt; i++)
867 fl->ifind = fl->ifcnt & 1;
870 readamt = read(fl->ifhandle, &fl->ifbuf[fl->ifind + fl->ifcnt], QUANTUM);
875 if ((fl->ifcnt += readamt) == 0)
884 int TokenizeLine(void)
886 char * ln = NULL; // Ptr to current position in line
887 char * p; // Random character ptr
888 TOKEN * tk; // Token-deposit ptr
889 int state = 0; // State for keyword detector
890 int j = 0; // Var for keyword detector
891 char c; // Random char
892 VALUE v; // Random value
893 char * nullspot = NULL; // Spot to clobber for SYMBOL terminatn
894 int stuffnull; // 1:terminate SYMBOL '\0' at *nullspot
896 int stringNum = 0; // Pointer to string locations in tokenized line
900 if (cur_inobj == NULL) // Return EOF if input stack is empty
903 // Get another line of input from the current input source: a file,
904 // a macro, or a repeat-block
905 switch (cur_inobj->in_type)
909 // o bump source line number;
910 // o tag the listing-line with a space;
911 // o kludge lines generated by Alcyon C.
913 if ((ln = GetNextLine()) == NULL)
915 if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IFILE...\n");
916 fpop(); // Pop input level
917 goto retry; // Try for more lines
920 curlineno++; // Bump line number
925 // AS68 compatibility, throw away all lines starting with
926 // back-quotes, tildes, or '*'
927 // On other lines, turn the first '*' into a semi-colon.
928 if (*ln == '`' || *ln == '~' || *ln == '*')
932 for(p=ln; *p!=EOS; p++)
945 // o Handle end-of-macro;
946 // o tag the listing-line with an at (@) sign.
948 if ((ln = GetNextMacroLine()) == NULL)
950 ExitMacro(); // Exit macro (pop args, do fpop(), etc)
951 goto retry; // Try for more lines...
957 // o Handle end-of-repeat-block;
958 // o tag the listing-line with a pound (#) sign.
960 if ((ln = GetNextRepeatLine()) == NULL)
962 if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n");
971 // Save text of the line. We only do this during listings and within
972 // macro-type blocks, since it is expensive to unconditionally copy every
977 // General house-keeping
978 tok = tokeol; // Set "tok" to EOL in case of error
979 tk = etok; // Reset token ptr
980 stuffnull = 0; // Don't stuff nulls
981 totlines++; // Bump total #lines assembled
983 // See if the entire line is a comment. This is a win if the programmer
984 // puts in lots of comments
985 if (*ln == '*' || *ln == ';' || ((*ln == '/') && (*(ln + 1) == '/')))
988 // Main tokenization loop;
989 // o skip whitespace;
990 // o handle end-of-line;
992 // o handle single-character tokens (operators, etc.);
993 // o handle multiple-character tokens (constants, strings, etc.).
996 // Skip whitespace, handle EOL
997 while ((int)chrtab[*ln] & WHITE)
1000 // Handle EOL, comment with ';'
1001 if (*ln == EOS || *ln == ';'|| ((*ln == '/') && (*(ln + 1) == '/')))
1004 // Handle start of symbol. Symbols are null-terminated in place. The
1005 // termination is always one symbol behind, since there may be no place
1006 // for a null in the case that an operator immediately follows the name.
1011 if (stuffnull) // Terminate old symbol from previous pass
1014 v = 0; // Assume no DOT attrib follows symbol
1016 p = nullspot = ln++; // Nullspot -> start of this symbol
1018 // Find end of symbol (and compute its length)
1019 for(j=1; (int)chrtab[*ln]&CTSYM; j++)
1022 // Handle "DOT" special forms (like ".b") that follow a normal
1023 // symbol or keyword:
1026 *ln++ = EOS; // Terminate symbol
1027 stuffnull = 0; // And never try it again
1029 // Character following the `.' must have a DOT attribute, and
1030 // the chararacter after THAT one must not have a start-symbol
1031 // attribute (to prevent symbols that look like, for example,
1032 // "zingo.barf", which might be a good idea anyway....)
1033 if ((((int)chrtab[*ln] & DOT) == 0) || ((int)dotxtab[*ln] <= 0))
1034 return error("[bwsl] must follow `.' in symbol");
1036 v = (VALUE)dotxtab[*ln++];
1038 if ((int)chrtab[*ln] & CTSYM)
1039 return error("misuse of `.', not allowed in symbols");
1042 // If the symbol is small, check to see if it's really the name of
1046 for(state=0; state>=0;)
1048 j = (int)tolowertab[*p++];
1051 if (kwcheck[j] != state)
1057 if (*p == EOS || p == ln)
1071 // Make j = -1 if user tries to use a RISC register while in 68K mode
1072 if (!(rgpu || rdsp) && ((TOKEN)j >= KW_R0 && (TOKEN)j <= KW_R31))
1077 //make j = -1 if time, date etc with no preceeding ^^
1078 //defined, referenced, streq, macdef, date and time
1081 case 112: // defined
1082 case 113: // referenced
1091 // If not tokenized keyword OR token was not found
1092 if (j < 0 || state < 0)
1096 //problem here: nullspot is a char * but TOKEN is a uint32_t. On a 64-bit system,
1097 //this will cause all kinds of mischief.
1099 *tk++ = (TOKEN)nullspot;
1101 string[stringNum] = nullspot;
1112 if (v) // Record attribute token (if any)
1115 if (stuffnull) // Arrange for string termination on next pass
1121 // Handle identity tokens
1128 // Handle multiple-character tokens
1133 case '!': // ! or !=
1143 case '\'': // 'string'
1144 case '\"': // "string"
1148 // More char * stuffing (8 bytes) into the space of 4 (TOKEN).
1149 // Need to figure out how to fix this crap.
1153 string[stringNum] = ln;
1158 for(p=ln; *ln!=EOS && *ln!=c1;)
1167 return(error("unterminated string"));
1196 warn("bad backslash code in string");
1206 return error("unterminated string");
1210 case '$': // $, hex constant
1211 if ((int)chrtab[*ln] & HDIGIT)
1215 while ((int)hextab[*ln] >= 0)
1216 v = (v << 4) + (int)hextab[*ln++];
1220 if ((*(ln + 1) == 'b') || (*(ln + 1) == 'B'))
1226 if ((*(ln + 1) == 'w') || (*(ln + 1) == 'W'))
1232 if ((*(ln + 1) == 'l') || (*(ln + 1) == 'L'))
1245 case '<': // < or << or <> or <=
1264 case ':': // : or ::
1274 case '=': // = or ==
1284 case '>': // > or >> or >=
1299 case '%': // % or binary constant
1300 if (*ln < '0' || *ln > '1')
1308 while (*ln >= '0' && *ln <= '1')
1309 v = (v << 1) + *ln++ - '0';
1313 if ((*(ln + 1) == 'b') || (*(ln + 1) == 'B'))
1319 if ((*(ln + 1) == 'w') || (*(ln + 1) == 'W'))
1325 if ((*(ln + 1) == 'l') || (*(ln + 1) == 'L'))
1334 case '@': // @ or octal constant
1335 if (*ln < '0' || *ln > '7')
1343 while (*ln >= '0' && *ln <= '7')
1344 v = (v << 3) + *ln++ - '0';
1348 if ((*(ln+1) == 'b') || (*(ln+1) == 'B'))
1354 if ((*(ln+1) == 'w') || (*(ln+1) == 'W'))
1360 if ((*(ln+1) == 'l') || (*(ln+1) == 'L'))
1369 case '^': // ^ or ^^ <operator-name>
1376 if (((int)chrtab[*++ln] & STSYM) == 0)
1378 error("invalid symbol following ^^");
1384 while ((int)chrtab[*ln] & CTSYM)
1387 for(state=0; state>=0;)
1389 // Get char, convert to lowercase
1392 if (j >= 'A' && j <= 'Z')
1397 if (kwcheck[j] != state)
1403 if (*p == EOS || p == ln)
1412 if (j < 0 || state < 0)
1414 error("unknown symbol following ^^");
1421 interror(2); // Bad MULTX entry in chrtab
1426 // Handle decimal constant
1431 while ((int)chrtab[*ln] & DIGIT)
1432 v = (v * 10) + *ln++ - '0';
1434 // See if there's a .[bwl] after the constant & deal with it if so
1437 if ((*(ln + 1) == 'b') || (*(ln + 1) == 'B'))
1442 else if ((*(ln + 1) == 'w') || (*(ln + 1) == 'W'))
1447 else if ((*(ln + 1) == 'l') || (*(ln + 1) == 'L'))
1455 //printf("CONST: %i\n", v);
1459 // Handle illegal character
1460 return error("illegal character");
1463 // Terminate line of tokens and return "success."
1466 tok = etok; // Set tok to beginning of line
1468 if (stuffnull) // Terminate last SYMBOL
1478 // .GOTO <label> goto directive
1480 // The label is searched for starting from the first line of the current,
1481 // enclosing macro definition. If no enclosing macro exists, an error is
1484 // A label is of the form:
1486 // :<name><whitespace>
1488 // The colon must appear in column 1. The label is stripped prior to macro
1489 // expansion, and is NOT subject to macro expansion. The whitespace may also
1492 //int d_goto(WORD siz) {
1494 int d_goto(WORD unused)
1496 // char * sym; // Label to search for
1497 // LONG * defln; // Macro definition strings
1498 char * s1; // Temps for string comparison
1500 // IMACRO * imacro; // Macro invocation block
1502 // Setup for the search
1504 return error("missing label");
1506 // sym = (char *)tok[1];
1507 char * sym = string[tok[1]];
1510 if (cur_inobj->in_type != SRC_IMACRO)
1511 return error("goto not in macro");
1513 IMACRO * imacro = cur_inobj->inobj.imacro;
1514 // defln = (LONG *)imacro->im_macro->svalue;
1515 struct LineList * defln = imacro->im_macro->lineList;
1517 // Find the label, starting with the first line.
1518 // for(; defln!=NULL; defln=(LONG *)*defln)
1519 for(; defln!=NULL; defln=defln->next)
1521 // if (*(char *)(defln + 1) == ':')
1522 if (defln->line[0] == ':')
1524 // Compare names (sleazo string compare)
1525 // This string compare is not right. Doesn't check for lengths.
1526 // (actually it does, but in a crappy, unclear way.)
1527 #warning "!!! Bad string comparison !!!"
1529 // s2 = (char *)(defln + 1) + 1;
1543 // Found the label, set new macro next-line and return.
1544 if ((*s2 == EOS) || ((int)chrtab[*s2] & WHITE))
1546 imacro->im_nextln = defln;
1552 return error("goto label not found");
1556 void DumpTokenBuffer(void)
1559 printf("Tokens [%X]: ", sloc);
1561 for(t=tokbuf; *t!=EOL; t++)
1565 else if (*t == CONST)
1568 printf("[CONST: $%X]", (uint32_t)*t);
1570 else if (*t == ACONST)
1572 else if (*t == STRING)
1573 // printf("[STRING]");
1576 printf("[STRING:\"%s\"]", string[*t]);
1578 else if (*t == SYMBOL)
1581 printf("[SYMBOL:\"%s\"]", string[*t]);
1585 else if (*t == TKEOF)
1587 else if (*t == DEQUALS)
1588 printf("[DEQUALS]");
1593 else if (*t == DCOLON)
1605 else if (*t == UNMINUS)
1606 printf("[UNMINUS]");
1607 else if (*t == DOTB)
1609 else if (*t == DOTW)
1611 else if (*t == DOTL)
1613 else if (*t == DOTI)
1615 else if (*t == ENDEXPR)
1616 printf("[ENDEXPR]");
1617 else if (*t == CR_DEFINED)
1618 printf("[CR_DEFINED]");
1619 else if (*t == CR_REFERENCED)
1620 printf("[CR_REFERENCED]");
1621 else if (*t == CR_STREQ)
1622 printf("[CR_STREQ]");
1623 else if (*t == CR_MACDEF)
1624 printf("[CR_MACDEF]");
1625 else if (*t == CR_TIME)
1626 printf("[CR_TIME]");
1627 else if (*t == CR_DATE)
1628 printf("[CR_DATE]");
1629 else if (*t >= 0x20 && *t <= 0x2F)
1630 printf("[%c]", (char)*t);
1631 else if (*t >= 0x3A && *t <= 0x3F)
1632 printf("[%c]", (char)*t);
1633 else if (*t >= 0x80 && *t <= 0x87)
1634 printf("[D%u]", ((uint32_t)*t) - 0x80);
1635 else if (*t >= 0x88 && *t <= 0x8F)
1636 printf("[A%u]", ((uint32_t)*t) - 0x88);
1638 printf("[%X:%c]", (uint32_t)*t, (char)*t);
1639 // printf("[%X]", (uint32_t)*t);