X-Git-Url: http://shamusworld.gotdns.org/cgi-bin/gitweb.cgi?p=rmac;a=blobdiff_plain;f=token.c;h=75b4cc166546e978932c1655faf1e3f70e2e2379;hp=c78d97ac6f0054d8ca02bf68b0a3a1c0c10406e4;hb=5cd8a4814b805f1ef8ce689423eb5eeba12573c5;hpb=689d9bc6a5776f54995b68ad2493652f55b9d419 diff --git a/token.c b/token.c index c78d97a..75b4cc1 100644 --- a/token.c +++ b/token.c @@ -1,30 +1,32 @@ // // RMAC - Reboot's Macro Assembler for the Atari Jaguar Console System // TOKEN.C - Token Handling -// Copyright (C) 199x Landon Dyer, 2011-2012 Reboot and Friends +// Copyright (C) 199x Landon Dyer, 2011-2017 Reboot and Friends // RMAC derived from MADMAC v1.07 Written by Landon Dyer, 1986 // Source utilised with the kind permission of Landon Dyer // #include "token.h" +#include "direct.h" #include "error.h" #include "macro.h" #include "procln.h" +#include "sect.h" #include "symbol.h" #define DECL_KW // Declare keyword arrays -#define DEF_KW // Declare keyword values +#define DEF_KW // Declare keyword values #include "kwtab.h" // Incl generated keyword tables & defs int lnsave; // 1; strcpy() text of current line -int curlineno; // Current line number +uint16_t curlineno; // Current line number (64K max currently) int totlines; // Total # of lines int mjump_align = 0; // mjump alignment flag char lntag; // Line tag char * curfname; // Current filename -char tolowertab[128]; // Uppercase ==> lowercase -char hextab[128]; // Table of hex values +char tolowertab[128]; // Uppercase ==> lowercase +int8_t hextab[128]; // Table of hex values char dotxtab[128]; // Table for ".b", ".s", etc. char irbuf[LNSIZ]; // Text for .rept block line char lnbuf[LNSIZ]; // Text of current line @@ -54,63 +56,87 @@ static IMACRO * f_imacro; // Ptr list of free IMACROs static TOKEN tokbuf[TOKBUFSIZE]; // Token buffer (stack-like, all files) char chrtab[] = { - ILLEG, ILLEG, ILLEG, ILLEG, // NUL SOH STX ETX - ILLEG, ILLEG, ILLEG, ILLEG, // EOT ENQ ACK BEL - ILLEG, WHITE, ILLEG, ILLEG, // BS HT LF VT - WHITE, ILLEG, ILLEG, ILLEG, // FF CR SO SI + ILLEG, ILLEG, ILLEG, ILLEG, // NUL SOH STX ETX + ILLEG, ILLEG, ILLEG, ILLEG, // EOT ENQ ACK BEL + ILLEG, WHITE, ILLEG, ILLEG, // BS HT LF VT + WHITE, ILLEG, ILLEG, ILLEG, // FF CR SO SI - ILLEG, ILLEG, ILLEG, ILLEG, // DLE DC1 DC2 DC3 - ILLEG, ILLEG, ILLEG, ILLEG, // DC4 NAK SYN ETB - ILLEG, ILLEG, ILLEG, ILLEG, // CAN EM SUB ESC - ILLEG, ILLEG, ILLEG, ILLEG, // FS GS RS US + ILLEG, ILLEG, ILLEG, ILLEG, // DLE DC1 DC2 DC3 + ILLEG, ILLEG, ILLEG, ILLEG, // DC4 NAK SYN ETB + ILLEG, ILLEG, ILLEG, ILLEG, // CAN EM SUB ESC + ILLEG, ILLEG, ILLEG, ILLEG, // FS GS RS US WHITE, MULTX, MULTX, SELF, // SP ! " # MULTX+CTSYM, MULTX, SELF, MULTX, // $ % & ' SELF, SELF, SELF, SELF, // ( ) * + SELF, SELF, STSYM, SELF, // , - . / - DIGIT+HDIGIT+CTSYM, DIGIT+HDIGIT+CTSYM, // 0 1 - DIGIT+HDIGIT+CTSYM, DIGIT+HDIGIT+CTSYM, // 2 3 - DIGIT+HDIGIT+CTSYM, DIGIT+HDIGIT+CTSYM, // 4 5 - DIGIT+HDIGIT+CTSYM, DIGIT+HDIGIT+CTSYM, // 6 7 - DIGIT+HDIGIT+CTSYM, DIGIT+HDIGIT+CTSYM, // 8 9 - MULTX, MULTX, // : ; - MULTX, MULTX, MULTX, STSYM+CTSYM, // < = > ? + DIGIT+HDIGIT+CTSYM, DIGIT+HDIGIT+CTSYM, // 0 1 + DIGIT+HDIGIT+CTSYM, DIGIT+HDIGIT+CTSYM, // 2 3 + DIGIT+HDIGIT+CTSYM, DIGIT+HDIGIT+CTSYM, // 4 5 + DIGIT+HDIGIT+CTSYM, DIGIT+HDIGIT+CTSYM, // 6 7 + DIGIT+HDIGIT+CTSYM, DIGIT+HDIGIT+CTSYM, // 8 9 + MULTX, MULTX, // : ; + MULTX, MULTX, MULTX, STSYM+CTSYM, // < = > ? MULTX, STSYM+CTSYM+HDIGIT, // @ A (char)((BYTE)DOT)+STSYM+CTSYM+HDIGIT, STSYM+CTSYM+HDIGIT, // B C - STSYM+CTSYM+HDIGIT, STSYM+CTSYM+HDIGIT, // D E + (char)((BYTE)DOT)+STSYM+CTSYM+HDIGIT, STSYM+CTSYM+HDIGIT, // D E STSYM+CTSYM+HDIGIT, STSYM+CTSYM, // F G - STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, // H I J K + STSYM+CTSYM, (char)((BYTE)DOT)+STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, // H I J K (char)((BYTE)DOT)+STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, // L M N O - STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, (char)((BYTE)DOT)+STSYM+CTSYM, // P Q R S + (char)((BYTE)DOT)+STSYM+CTSYM, (char)((BYTE)DOT)+STSYM+CTSYM, STSYM+CTSYM, (char)((BYTE)DOT)+STSYM+CTSYM, // P Q R S STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, (char)((BYTE)DOT)+STSYM+CTSYM, // T U V W STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, SELF, // X Y Z [ SELF, SELF, MULTX, STSYM+CTSYM, // \ ] ^ _ ILLEG, STSYM+CTSYM+HDIGIT, // ` a (char)((BYTE)DOT)+STSYM+CTSYM+HDIGIT, STSYM+CTSYM+HDIGIT, // b c - STSYM+CTSYM+HDIGIT, STSYM+CTSYM+HDIGIT, // d e + (char)((BYTE)DOT)+STSYM+CTSYM+HDIGIT, STSYM+CTSYM+HDIGIT, // d e STSYM+CTSYM+HDIGIT, STSYM+CTSYM, // f g - STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, // h i j k + STSYM+CTSYM, (char)((BYTE)DOT)+STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, // h i j k (char)((BYTE)DOT)+STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, // l m n o - STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, (char)((BYTE)DOT)+STSYM+CTSYM, // p q r s + (char)((BYTE)DOT)+STSYM+CTSYM, (char)((BYTE)DOT)+STSYM+CTSYM, STSYM+CTSYM, (char)((BYTE)DOT)+STSYM+CTSYM, // p q r s STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, (char)((BYTE)DOT)+STSYM+CTSYM, // t u v w - STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, SELF, // x y z { - SELF, SELF, SELF, ILLEG // | } ~ DEL + (char)((BYTE)DOT)+STSYM+CTSYM, STSYM+CTSYM, STSYM+CTSYM, SELF, // x y z { + SELF, SELF, SELF, ILLEG // | } ~ DEL }; // Names of registers static char * regname[] = { - "d0", "d1", "d2", "d3", "d4", "d5", "d6", "d7", - "a0", "a1", "a2", "a3", "a4", "a5", "a6", "a7", - "pc", "ssp", "usp", "sr", "ccr" +// "d0", "d1", "d2", "d3", "d4", "d5", "d6", "d7", +// "a0", "a1", "a2", "a3", "a4", "a5", "a6", "a7", +// "pc", "ssp", "usp", "sr", "ccr" + "d0","d1","d2","d3","d4","d5","d6","d7", // 128,135 + "a0","a1","a2","a3","a4","a5","a6","sp", // 136,143 + "ssp","pc","sr","ccr","regequ","set","reg","r0", // 144,151 + "r1","r2","r3","r4","r5","r6","r7","r8", // 152,159 + "r9","r10","r11","r12","r13","r14","r15","r16", // 160,167 + "r17","r18","r19","r20","r21","r22","r23","r24", // 168,175 + "r25","r26","r27","r28","r29","r30","r31","ccdef", // 176,183 + "usp","ic40","dc40","bc40","sfc","dfc","","vbr", // 184,191 + "cacr","caar","msp","isp","tc","itt0","itt1","dtt0", // 192,199 + "dtt1","mmusr","urp","srp","iacr0","iacr1","dacr0","dacr1", // 200,207 + "tt0","tt1","crp","","","","","", // 208,215 + "","","","","fpiar","fpsr","fpcr","", // 216,223 + "fp0","fp1","fp2","fp3","fp4","fp5","fp6","fp7", // 224,231 + "","","","","","","","", // 232,239 + "","","","","","","","", // 240,247 + "","","","","","","","", // 248,255 + "","","","","x0","x1","y0","y1", // 256,263 + "","b0","","b2","","b1","a","b", // 264,271 + "mr","omr","la","lc","ssh","ssl","ss","", // 272,279 + "n0","n1","n2","n3","n4","n5","n6","n7", // 280,287 + "m0","m1","m2","m3","m4","m5","m6","m7", // 288,295 + "","","","","","","l","p", // 296,303 + "mr","omr","la","lc","ssh","ssl","ss","", // 304,311 + "a10","b10","x","y","","","ab","ba" // 312,319 }; static char * riscregname[] = { - "r0", "r1", "r2", "r3", "r4", "r5", "r6", "r7", + "r0", "r1", "r2", "r3", "r4", "r5", "r6", "r7", "r8", "r9", "r10", "r11", "r12", "r13", "r14", "r15", "r16", "r17", "r18", "r19", "r20", "r21", "r22", "r23", "r24", "r25", "r26", "r27", "r28", "r29", "r30", "r31" @@ -155,16 +181,26 @@ void InitTokenizer(void) tolowertab[i] |= 0x20; // These characters are legal immediately after a period - dotxtab['b'] = DOTB; // .b .B .s .S + dotxtab['b'] = DOTB; // .b .B .s .S dotxtab['B'] = DOTB; - dotxtab['s'] = DOTB; - dotxtab['S'] = DOTB; + //dotxtab['s'] = DOTB; + //dotxtab['S'] = DOTB; dotxtab['w'] = DOTW; // .w .W dotxtab['W'] = DOTW; dotxtab['l'] = DOTL; // .l .L dotxtab['L'] = DOTL; dotxtab['i'] = DOTI; // .i .I (???) dotxtab['I'] = DOTI; + dotxtab['D'] = DOTD; // .d .D (quad word) + dotxtab['d'] = DOTD; + dotxtab['S'] = DOTS; // .s .S + dotxtab['s'] = DOTS; + dotxtab['Q'] = DOTQ; // .q .Q + dotxtab['q'] = DOTQ; + dotxtab['X'] = DOTX; // .x .x + dotxtab['x'] = DOTX; + dotxtab['P'] = DOTP; // .p .P + dotxtab['p'] = DOTP; } @@ -230,7 +266,7 @@ INOBJ * a_inobj(int typ) inobj->inobj.ifile = ifile; break; - case SRC_IMACRO: // Alloc and init an IMACRO + case SRC_IMACRO: // Alloc and init an IMACRO if (f_imacro == NULL) imacro = malloc(sizeof(IMACRO)); else @@ -288,7 +324,6 @@ int ExpandMacro(char * src, char * dest, int destsiz) IMACRO * imacro = cur_inobj->inobj.imacro; int macnum = (int)(imacro->im_macro->sattr); -// destsiz--; char * dst = dest; // Next dest slot char * edst = dest + destsiz - 1; // End + 1(?) of dest buffer @@ -314,6 +349,11 @@ int ExpandMacro(char * src, char * dest, int destsiz) if (dst >= edst) goto overflow; + // Skip comments in case a loose @ or \ is in there + // In that case the tokeniser was trying to expand it. + if ((*s == ';') || ((*s == '/') && (*(s + 1) == '/'))) + goto skipcomments; + *dst++ = *s++; } // Do macro expansion @@ -330,11 +370,11 @@ int ExpandMacro(char * src, char * dest, int destsiz) *dst++ = *s++; continue; - case '?': // \? set `questmark' flag + case '?': // \? set `questmark' flag ++s; questmark = 1; break; - case '#': // \#, number of arguments + case '#': // \#, number of arguments sprintf(numbuf, "%d", (int)imacro->im_nargs); goto copystr; case '!': // \! size suffix supplied on invocation @@ -347,7 +387,7 @@ int ExpandMacro(char * src, char * dest, int destsiz) } goto copy_d; - case '~': // ==> unique label string Mnnnn... + case '~': // ==> unique label string Mnnnn... sprintf(numbuf, "M%u", curuniq); copystr: d = numbuf; @@ -511,11 +551,11 @@ DEBUG printf("ExM: SYMBOL=\"%s\"", d); *dst++ = '"'; continue; break; -// Shamus: Changing the format specifier from %lx to %ux caused -// the assembler to choke on legitimate code... Need to investigate -// this further before changing anything else here! +// Shamus: Changing the format specifier from %lx to %ux caused the assembler +// to choke on legitimate code... Need to investigate this further +// before changing anything else here! case CONST: - sprintf(numbuf, "$%lx", (LONG)*tk++); + sprintf(numbuf, "$%lx", (long unsigned int)*tk++); d = numbuf; break; case DEQUALS: @@ -554,6 +594,9 @@ DEBUG printf("ExM: SYMBOL=\"%s\"", d); case DOTL: d = ".l"; break; + case CR_ABSCOUNT: + d = "^^abscount"; + break; case CR_DATE: d = "^^date"; break; @@ -601,6 +644,8 @@ strcopy: } } +skipcomments: + *dst = EOS; DEBUG { printf("ExM: dst=\"%s\"\n", dest); } return OK; @@ -617,8 +662,6 @@ overflow: // char * GetNextMacroLine(void) { -// unsigned source_addr; - IMACRO * imacro = cur_inobj->inobj.imacro; // LONG * strp = imacro->im_nextln; struct LineList * strp = imacro->im_nextln; @@ -626,7 +669,6 @@ char * GetNextMacroLine(void) if (strp == NULL) // End-of-macro return NULL; -// imacro->im_nextln = (LONG *)*strp; imacro->im_nextln = strp->next; // ExpandMacro((char *)(strp + 1), imacro->im_lnbuf, LNSIZ); ExpandMacro(strp->line, imacro->im_lnbuf, LNSIZ); @@ -640,7 +682,6 @@ char * GetNextMacroLine(void) // char * GetNextRepeatLine(void) { - IREPT * irept = cur_inobj->inobj.irept; LONG * strp = irept->ir_nextln; // initial null @@ -672,17 +713,13 @@ char * GetNextRepeatLine(void) // int include(int handle, char * fname) { - IFILE * ifile; - INOBJ * inobj; - FILEREC * fr; - - // Verbose mode - if (verb_flag) + // Debug mode + if (debug) printf("[include: %s, cfileno=%u]\n", fname, cfileno); // Alloc and initialize include-descriptors - inobj = a_inobj(SRC_IFILE); - ifile = inobj->inobj.ifile; + INOBJ * inobj = a_inobj(SRC_IFILE); + IFILE * ifile = inobj->inobj.ifile; ifile->ifhandle = handle; // Setup file handle ifile->ifind = ifile->ifcnt = 0; // Setup buffer indices @@ -690,21 +727,20 @@ int include(int handle, char * fname) ifile->ifoldfname = curfname; // Save old filename ifile->ifno = cfileno; // Save old file number -// cfileno = filecount++; // Compute new file number // NB: This *must* be preincrement, we're adding one to the filecount here! cfileno = ++filecount; // Compute NEW file number curfname = strdup(fname); // Set current filename (alloc storage) curlineno = 0; // Start on line zero // Add another file to the file-record - fr = (FILEREC *)malloc(sizeof(FILEREC)); + FILEREC * fr = (FILEREC *)malloc(sizeof(FILEREC)); fr->frec_next = NULL; fr->frec_name = curfname; if (last_fr == NULL) - filerec = fr; // Add first filerec + filerec = fr; // Add first filerec else - last_fr->frec_next = fr; // Append to list of filerecs + last_fr->frec_next = fr; // Append to list of filerecs last_fr = fr; DEBUG printf("[include: curfname: %s, cfileno=%u]\n", curfname, cfileno); @@ -727,33 +763,41 @@ int fpop(void) { // Pop IFENT levels until we reach the conditional assembly context we // were at when the input object was entered. + int numUnmatched = 0; + while (ifent != inobj->in_ifent) { - if (d_endif() != 0) // Something bad happened during endif parsing? - return -1; // If yes, bail instead of getting stuck in a loop + if (d_endif() != 0) // Something bad happened during endif parsing? + return -1; // If yes, bail instead of getting stuck in a loop + + numUnmatched++; } - tok = inobj->in_otok; // Restore tok and otok + // Give a warning to the user that we had to wipe their bum for them + if (numUnmatched > 0) + warni("missing %d .endif(s)", numUnmatched); + + tok = inobj->in_otok; // Restore tok and otok etok = inobj->in_etok; switch (inobj->in_type) { - case SRC_IFILE: // Pop and release an IFILE - if (verb_flag) + case SRC_IFILE: // Pop and release an IFILE + if (debug) printf("[Leaving: %s]\n", curfname); ifile = inobj->inobj.ifile; ifile->if_link = f_ifile; f_ifile = ifile; close(ifile->ifhandle); // Close source file -if (verb_flag) printf("[fpop (pre): curfname=%s]\n", curfname); +if (debug) printf("[fpop (pre): curfname=%s]\n", curfname); curfname = ifile->ifoldfname; // Set current filename -if (verb_flag) printf("[fpop (post): curfname=%s]\n", curfname); -if (verb_flag) printf("[fpop: (pre) cfileno=%d ifile->ifno=%d]\n", (int)cfileno, (int)ifile->ifno); - curlineno = ifile->ifoldlineno; // Set current line# +if (debug) printf("[fpop (post): curfname=%s]\n", curfname); +if (debug) printf("[fpop: (pre) cfileno=%d ifile->ifno=%d]\n", (int)cfileno, (int)ifile->ifno); + curlineno = ifile->ifoldlineno; // Set current line# DEBUG printf("cfileno=%d ifile->ifno=%d\n", (int)cfileno, (int)ifile->ifno); cfileno = ifile->ifno; // Restore current file number -if (verb_flag) printf("[fpop: (post) cfileno=%d ifile->ifno=%d]\n", (int)cfileno, (int)ifile->ifno); +if (debug) printf("[fpop: (post) cfileno=%d ifile->ifno=%d]\n", (int)cfileno, (int)ifile->ifno); break; case SRC_IMACRO: // Pop and release an IMACRO imacro = inobj->inobj.imacro; @@ -809,7 +853,7 @@ char * GetNextLine(void) if (*p == '\r') { if (i >= j) - break; // Need to read more, then look for '\n' to eat + break; // Need to read more, then look for '\n' to eat else if (p[1] == '\n') i++; } @@ -908,9 +952,14 @@ retry: case SRC_IFILE: if ((ln = GetNextLine()) == NULL) { -if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IFILE...\n"); - fpop(); // Pop input level - goto retry; // Try for more lines +if (debug) printf("TokenizeLine: Calling fpop() from SRC_IFILE...\n"); + if (fpop() == 0) // Pop input level + goto retry; // Try for more lines + else + { + ifent->if_prev = (IFENT *) - 1; //Signal Assemble() that we have reached EOF with unbalanced if/endifs + return TKEOF; + } } curlineno++; // Bump line number @@ -957,7 +1006,7 @@ if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IFILE...\n"); case SRC_IREPT: if ((ln = GetNextRepeatLine()) == NULL) { -if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n"); +if (debug) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n"); fpop(); goto retry; } @@ -996,7 +1045,7 @@ if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n"); ln++; // Handle EOL, comment with ';' - if (*ln == EOS || *ln == ';'|| ((*ln == '/') && (*(ln + 1) == '/'))) + if (*ln == EOS || *ln == ';'|| ((*ln == '/') && (*(ln + 1) == '/'))) break; // Handle start of symbol. Symbols are null-terminated in place. The @@ -1011,6 +1060,27 @@ if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n"); v = 0; // Assume no DOT attrib follows symbol stuffnull = 1; + + // In some cases, we need to check for a DOTx at the *beginning* + // of a symbol, as the "start" of the line we're currently looking + // at could be somewhere in the middle of that line! + if (*ln == '.') + { + // Make sure that it's *only* a .[bwsl] following, and not the + // start of a local symbol: + if ((chrtab[*(ln + 1)] & DOT) + && (dotxtab[*(ln + 1)] != 0) + && !(chrtab[*(ln + 2)] & CTSYM)) + { + // We found a legitimate DOTx construct, so add it to the + // token stream: + ln++; + stuffnull = 0; + *tk++ = (TOKEN)dotxtab[*ln++]; + continue; + } + } + p = nullspot = ln++; // Nullspot -> start of this symbol // Find end of symbol (and compute its length) @@ -1022,19 +1092,19 @@ if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n"); if (*ln == '.') { *ln++ = EOS; // Terminate symbol - stuffnull = 0; // And never try it again + stuffnull = 0; // And never try it again // Character following the `.' must have a DOT attribute, and // the chararacter after THAT one must not have a start-symbol // attribute (to prevent symbols that look like, for example, // "zingo.barf", which might be a good idea anyway....) - if ((((int)chrtab[*ln] & DOT) == 0) || ((int)dotxtab[*ln] <= 0)) - return error("[bwsl] must follow `.' in symbol"); + if (((chrtab[*ln] & DOT) == 0) || (dotxtab[*ln] == 0)) + return error("[bwsl] must follow '.' in symbol"); v = (VALUE)dotxtab[*ln++]; - if ((int)chrtab[*ln] & CTSYM) - return error("misuse of `.', not allowed in symbols"); + if (chrtab[*ln] & CTSYM) + return error("misuse of '.', not allowed in symbols"); } // If the symbol is small, check to see if it's really the name of @@ -1086,7 +1156,7 @@ if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n"); } // If not tokenized keyword OR token was not found - if (j < 0 || state < 0) + if ((j < 0) || (state < 0)) { *tk++ = SYMBOL; //#warning @@ -1125,9 +1195,10 @@ if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n"); // Handle multiple-character tokens if (c & MULTX) { + switch (*ln++) { - case '!': // ! or != + case '!': // ! or != if (*ln == '=') { *tk++ = NE; @@ -1137,20 +1208,21 @@ if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n"); *tk++ = '!'; continue; - case '\'': // 'string' - case '\"': // "string" - c1 = ln[-1]; + case '\'': // 'string' + if (m6502) + { + // Hardcoded for now, maybe this will change in the future + *tk++ = STRINGA8; + goto dostring; + } + // Fall through + case '\"': // "string" *tk++ = STRING; -//#warning -// More char * stuffing (8 bytes) into the space of 4 (TOKEN). -// Need to figure out how to fix this crap. -#if 0 - *tk++ = (TOKEN)ln; -#else +dostring: + c1 = ln[-1]; string[stringNum] = ln; *tk++ = stringNum; stringNum++; -#endif for(p=ln; *ln!=EOS && *ln!=c1;) { @@ -1205,50 +1277,17 @@ if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n"); *p++ = EOS; continue; case '$': // $, hex constant - if ((int)chrtab[*ln] & HDIGIT) + if (chrtab[*ln] & HDIGIT) { v = 0; // Parse the hex value - while ((int)hextab[*ln] >= 0) + while (hextab[*ln] >= 0) v = (v << 4) + (int)hextab[*ln++]; - // ggn: Okay, some comments here are in order I think.... - // The original madmac sources didn't parse the size at - // this point (i.e. .b/.w/.l). It was probably done at - // another point, although it's unclear to me exactly - // where. So why change this? My understanding (at least - // from what SCPCD said on IRC) is that .w addressing - // formats produce wrong code on jaguar (or doesn't execute - // properly? something like that). So the code was changed - // to mask off the upper bits depending on length (note: I - // don't think .b is valid at all! I only know of .w/.l, so - // this should probably be wiped). Then the code that - // parses the constant and checks to see if it's between - // $ffff0000 and $8000 never got triggered, so yay job - // done! ...now say we want to assemble a st .prg. One of - // the most widely spread optimisations is move.X expr.w,Y - // (or vice versa, or both, anyway...) to access hardware - // registers (which are mapped to $fxxxxx). This botchy - // thing would create "hilarious" code while trying to - // access hardware registers. So I made a condition to see - // if st mode or jaguar is active and apply the both or - // not. One last note: this is hardcoded to get optimised - // for now on ST mode, i.e. it can't generate code like - // move.w $00001234,d0 - it'll always get optimised to - // move.w $1234.w,d0. It's probably ok, but maybe a warning - // should be emitted? Or maybe finding a way to make it not - // auto-optimise? I think it's ok for now... if (*ln == '.') { - if (obj_format == ALCYON) - { - if ((*(ln + 1) == 'b') || (*(ln + 1) == 'B') || (*(ln + 1) == 'w') || (*(ln + 1) == 'W') || (*(ln + 1) == 'l') || (*(ln + 1) == 'L')) - { - ln += 2; - } - } - else + if (obj_format == BSD) { if ((*(ln + 1) & 0xDF) == 'B') { @@ -1269,12 +1308,29 @@ if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n"); *tk++ = CONST; *tk++ = v; + + if (obj_format == ALCYON) + { + if (*ln == '.') + { + if ((*(ln + 1) == 'w') || (*(ln + 1) == 'W')) + { + *tk++ = DOTW; + ln += 2; + } + else if ((*(ln + 1) == 'l') || (*(ln + 1) == 'L')) + { + *tk++ = DOTL; + ln += 2; + } + } + } } else *tk++ = '$'; continue; - case '<': // < or << or <> or <= + case '<': // < or << or <> or <= switch (*ln) { case '<': @@ -1303,7 +1359,7 @@ if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n"); *tk++ = ':'; continue; - case '=': // = or == + case '=': // = or == if (*ln == '=') { *tk++ = DEQUALS; @@ -1313,7 +1369,7 @@ if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n"); *tk++ = '='; continue; - case '>': // > or >> or >= + case '>': // > or >> or >= switch (*ln) { case '>': @@ -1328,7 +1384,7 @@ if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n"); *tk++ = '>'; continue; } - case '%': // % or binary constant + case '%': // % or binary constant if (*ln < '0' || *ln > '1') { *tk++ = '%'; @@ -1363,7 +1419,7 @@ if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n"); *tk++ = CONST; *tk++ = v; continue; - case '@': // @ or octal constant + case '@': // @ or octal constant if (*ln < '0' || *ln > '7') { *tk++ = '@'; @@ -1418,7 +1474,7 @@ if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n"); for(state=0; state>=0;) { - // Get char, convert to lowercase + // Get char, convert to lowercase j = *p++; if (j >= 'A' && j <= 'Z') @@ -1508,30 +1564,25 @@ goteol: // // .GOTO