X-Git-Url: http://shamusworld.gotdns.org/cgi-bin/gitweb.cgi?p=rmac;a=blobdiff_plain;f=token.c;h=e35e9bfc787697ea21e767f4a94397fb7ba223bb;hp=4f812c234725e81ec527dda3af041d653a750471;hb=75969398d9b8a9f82ea76fc4e4cbfb97b11160a4;hpb=3385b366632d03745033fa6b19faabf60219bc6b diff --git a/token.c b/token.c index 4f812c2..e35e9bf 100644 --- a/token.c +++ b/token.c @@ -7,31 +7,33 @@ // #include "token.h" -#include "symbol.h" -#include "procln.h" -#include "macro.h" #include "error.h" +#include "macro.h" +#include "procln.h" +#include "symbol.h" -#define DECL_KW // Declare keyword arrays -#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 -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 dotxtab[128]; // Table for ".b", ".s", etc. -char irbuf[LNSIZ]; // Text for .rept block line -char lnbuf[LNSIZ]; // Text of current line -WORD filecount; // Unique file number counter -WORD cfileno; // Current file number -TOKEN * tok; // Ptr to current token -TOKEN * etok; // Ptr past last token in tokbuf[] -TOKEN tokeol[1] = {EOL}; // Bailout end-of-line token +#define DECL_KW // Declare keyword arrays +#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 +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 dotxtab[128]; // Table for ".b", ".s", etc. +char irbuf[LNSIZ]; // Text for .rept block line +char lnbuf[LNSIZ]; // Text of current line +WORD filecount; // Unique file number counter +WORD cfileno; // Current file number +TOKEN * tok; // Ptr to current token +TOKEN * etok; // Ptr past last token in tokbuf[] +TOKEN tokeol[1] = {EOL}; // Bailout end-of-line token +char * string[TOKBUFSIZE*2]; // Token buffer string pointer storage // File record, used to maintain a list of every include file ever visited #define FILEREC struct _filerec @@ -44,90 +46,156 @@ FILEREC FILEREC * filerec; FILEREC * last_fr; -INOBJ * cur_inobj; // Ptr current input obj (IFILE/IMACRO) -static INOBJ * f_inobj; // Ptr list of free INOBJs -static IFILE * f_ifile; // Ptr list of free IFILEs -static IMACRO * f_imacro; // Ptr list of free IMACROs +INOBJ * cur_inobj; // Ptr current input obj (IFILE/IMACRO) +static INOBJ * f_inobj; // Ptr list of free INOBJs +static IFILE * f_ifile; // Ptr list of free IFILEs +static IMACRO * f_imacro; // Ptr list of free IMACROs -static TOKEN tokbuf[TOKBUFSIZE]; // Token buffer (stack-like, all files) +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, // 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, // < = > ? - - MULTX, STSYM+CTSYM+HDIGIT, // @ A - (char)((BYTE)DOT)+STSYM+CTSYM+HDIGIT, STSYM+CTSYM+HDIGIT, // B C - 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 - (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 - 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 - STSYM+CTSYM+HDIGIT, STSYM+CTSYM, // f g - STSYM+CTSYM, 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 - 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 + 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 + + 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, // < = > ? + + MULTX, STSYM+CTSYM+HDIGIT, // @ A + (char)((BYTE)DOT)+STSYM+CTSYM+HDIGIT, STSYM+CTSYM+HDIGIT, // B C + 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 + (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 + 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 + STSYM+CTSYM+HDIGIT, STSYM+CTSYM, // f g + STSYM+CTSYM, 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 + 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 }; // 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" }; static char * riscregname[] = { - "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" + "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" }; // -// Make `fnum' the Current `curfname' +// Initialize Tokenizer // -void setfnum(WORD fnum) +void InitTokenizer(void) { - FILEREC * fr; + int i; // Iterator + char * htab = "0123456789abcdefABCDEF"; // Hex character table + + lnsave = 0; // Don't save lines + curfname = ""; // No file, empty filename + filecount = (WORD)-1; + cfileno = (WORD)-1; // cfileno gets bumped to 0 + curlineno = 0; + totlines = 0; + etok = tokbuf; + f_inobj = NULL; + f_ifile = NULL; + f_imacro = NULL; + cur_inobj = NULL; + filerec = NULL; + last_fr = NULL; + lntag = SPACE; + + // Initialize hex, "dot" and tolower tables + for(i=0; i<128; i++) + { + hextab[i] = -1; + dotxtab[i] = 0; + tolowertab[i] = (char)i; + } - for(fr=filerec; fr!=NULL && fnum--; fr=fr->frec_next); + for(i=0; htab[i]!=EOS; i++) + hextab[htab[i]] = (char)((i < 16) ? i : i - 6); - if (fr == NULL) + for(i='A'; i<='Z'; i++) + tolowertab[i] |= 0x20; + + // These characters are legal immediately after a period + dotxtab['b'] = DOTB; // .b .B .s .S + dotxtab['B'] = 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; +} + + +void SetFilenameForErrorReporting(void) +{ + WORD fnum = cfileno; + + // Check for absolute top filename (this should never happen) + if (fnum == -1) + { curfname = "(*top*)"; - else - curfname = fr->frec_name; + return; + } + + FILEREC * fr = filerec; + + // Advance to the correct record... + while (fr != NULL && fnum != 0) + { + fr = fr->frec_next; + fnum--; + } + + // Check for file # record not found (this should never happen either) + if (fr == NULL) + { + curfname = "(*NOT FOUND*)"; + return; + } + + curfname = fr->frec_name; } @@ -142,7 +210,7 @@ INOBJ * a_inobj(int typ) // Allocate and initialize INOBJ first if (f_inobj == NULL) - inobj = (INOBJ *)amem((LONG)sizeof(INOBJ)); + inobj = malloc(sizeof(INOBJ)); else { inobj = f_inobj; @@ -151,9 +219,9 @@ INOBJ * a_inobj(int typ) switch (typ) { - case SRC_IFILE: // Alloc and init an IFILE + case SRC_IFILE: // Alloc and init an IFILE if (f_ifile == NULL) - ifile = (IFILE *)amem((LONG)sizeof(IFILE)); + ifile = malloc(sizeof(IFILE)); else { ifile = f_ifile; @@ -162,9 +230,9 @@ 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 = (IMACRO *)amem((LONG)sizeof(IMACRO)); + imacro = malloc(sizeof(IMACRO)); else { imacro = f_imacro; @@ -173,14 +241,14 @@ INOBJ * a_inobj(int typ) inobj->inobj.imacro = imacro; break; - case SRC_IREPT: // Alloc and init an IREPT - inobj->inobj.irept = (IREPT *)amem((LONG)sizeof(IREPT)); + case SRC_IREPT: // Alloc and init an IREPT + inobj->inobj.irept = malloc(sizeof(IREPT)); DEBUG printf("alloc IREPT\n"); break; } // Install INOBJ on top of input stack - inobj->in_ifent = ifent; // Record .if context on entry + inobj->in_ifent = ifent; // Record .if context on entry inobj->in_type = (WORD)typ; inobj->in_otok = tok; inobj->in_etok = etok; @@ -205,70 +273,71 @@ INOBJ * a_inobj(int typ) // (the colon must be in the first column). These labels are stripped before // macro expansion takes place. // -int mexpand(char * src, char * dest, int destsiz) +int ExpandMacro(char * src, char * dest, int destsiz) { - char * s; - char * d = NULL; - char * dst; // Next dest slot - char * edst; // End+1 of dest buffer int i; - int questmark; // \? for testing argument existence + int questmark; // \? for testing argument existence + char mname[128]; // Assume max size of a formal arg name + char numbuf[20]; // Buffer for text of CONSTs TOKEN * tk; - char mname[128]; // Assume max size of a formal arg name - int macnum; SYM * arg; - IMACRO * imacro; - char numbuf[20]; // Buffer for text of CONSTs + char ** symbolString; - imacro = cur_inobj->inobj.imacro; - macnum = (int)(imacro->im_macro->sattr); + DEBUG { printf("ExM: src=\"%s\"\n", src); } - --destsiz; - dst = dest; - edst = dest + 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 // Check for (and skip over) any "label" on the line - s = src; + char * s = src; + char * d = NULL; + if (*s == ':') { while (*s != EOS && !(chrtab[*s] & WHITE)) - ++s; + s++; if (*s != EOS) - ++s; // Skip first whitespace + s++; // Skip first whitespace } // Expand the rest of the line while (*s != EOS) { + // Copy single character if (*s != '\\') - { // Copy single character + { if (dst >= edst) goto overflow; *dst++ = *s++; } + // Do macro expansion else - { // Do macro expansion + { questmark = 0; // Do special cases switch (*++s) { - case '\\': // \\, \ (collapse to single backslash) + case '\\': // \\, \ (collapse to single backslash) if (dst >= edst) goto overflow; *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 + case '!': // \! size suffix supplied on invocation switch ((int)imacro->im_siz) { case SIZN: d = ""; break; @@ -278,12 +347,12 @@ int mexpand(char * src, char * dest, int destsiz) } goto copy_d; - case '~': // ==> unique label string Mnnnn... - sprintf(numbuf, "M%ud", curuniq); + case '~': // ==> unique label string Mnnnn... + sprintf(numbuf, "M%u", curuniq); copystr: d = numbuf; copy_d: - ++s; + s++; while (*d != EOS) { @@ -311,44 +380,62 @@ copy_d: // Get argument name: \name, \{name} d = mname; + + // \label if (*s != '{') - { // \foo + { do { *d++ = *s++; } while (chrtab[*s] & CTSYM); } + // \\{label} else - { // \\{foo} + { for(++s; *s != EOS && *s != '}';) *d++ = *s++; if (*s != '}') return error("missing '}'"); else - ++s; + s++; } *d = EOS; - // Lookup the argument and copy its (string) value into the destination string - DEBUG printf("mname='%s'\n", mname); + // Lookup the argument and copy its (string) value into the + // destination string + DEBUG printf("argument='%s'\n", mname); if ((arg = lookup(mname, MACARG, macnum)) == NULL) return errors("undefined argument: '%s'", mname); else { - // Convert a string of tokens (terminated with EOL) back into text. If an argument - // is out of range (not specified in the macro invocation) then it is ignored. + // Convert a string of tokens (terminated with EOL) back into + // text. If an argument is out of range (not specified in the + // macro invocation) then it is ignored. i = (int)arg->svalue; arg_num: - DEBUG printf("~argnumber=%d\n", i); - + DEBUG printf("~argnumber=%d (argBase=%u)\n", i, imacro->argBase); tk = NULL; if (i < imacro->im_nargs) - tk = argp[i]; + { +#if 0 +// tk = argp[i]; +// tk = argPtrs[i]; + tk = argPtrs[imacro->argBase + i]; +#else + tk = imacro->argument[i].token; + symbolString = imacro->argument[i].string; +//DEBUG +//{ +// printf("ExM: Preparing to parse argument #%u...\n", i); +// dumptok(tk); +//} +#endif + } // \?arg yields: // 0 if the argument is empty or non-existant, @@ -365,12 +452,13 @@ arg_num: continue; } - if (tk != NULL) // arg# is in range, so expand it + // Argument # is in range, so expand it + if (tk != NULL) { while (*tk != EOL) { - // Reverse-translation from a token number to a string. This is a hack. - // It might be better table-driven. + // Reverse-translation from a token number to a string. + // This is a hack. It might be better table-driven. d = NULL; if ((*tk >= KW_D0) && !rdsp && !rgpu) @@ -388,11 +476,22 @@ arg_num: switch ((int)*tk++) { case SYMBOL: - d = (char *)*tk++; +#if 0 +// d = (char *)*tk++; + d = string[*tk++]; +#else + // This fix should be done for strings too + d = symbolString[*tk++]; +DEBUG printf("ExM: SYMBOL=\"%s\"", d); +#endif break; case STRING: - d = (char *)*tk++; - +#if 0 +// d = (char *)*tk++; + d = string[*tk++]; +#else + d = symbolString[*tk++]; +#endif if (dst >= edst) goto overflow; @@ -477,7 +576,7 @@ arg_num: if (dst >= edst) goto overflow; - *dst++ = (char)*(tk-1); + *dst++ = (char)*(tk - 1); break; } } @@ -503,10 +602,12 @@ strcopy: } *dst = EOS; + DEBUG { printf("ExM: dst=\"%s\"\n", dest); } return OK; overflow: *dst = EOS; + DEBUG printf("*** OVERFLOW LINE ***\n%s\n", dest); return fatal("line too long as a result of macro expansion"); } @@ -514,42 +615,21 @@ overflow: // // Get Next Line of Text from a Macro // -char * getmln(void) +char * GetNextMacroLine(void) { - IMACRO * imacro; - LONG * strp; unsigned source_addr; - imacro = cur_inobj->inobj.imacro; - strp = imacro->im_nextln; + IMACRO * imacro = cur_inobj->inobj.imacro; +// LONG * strp = imacro->im_nextln; + struct LineList * strp = imacro->im_nextln; - if (strp == NULL) // End-of-macro + if (strp == NULL) // End-of-macro return NULL; - imacro->im_nextln = (LONG *)*strp; - mexpand((char *)(strp + 1), imacro->im_lnbuf, LNSIZ); - - if (!strcmp(imacro->im_macro->sname, "mjump") && !mjump_align) - { - // if we need to adjust the alignment of the jump source address to meet the rules of - // gpu main execution we need to skip the first nop of the macro. This is simpler than - // trying to insert nop's mid macro. - source_addr = (orgactive) ? orgaddr : sloc; - source_addr += 8; - - if (source_addr % 4) - { - strp = imacro->im_nextln; - - if (strp == NULL) - return NULL; - - imacro->im_nextln = (LONG *)*strp; - mexpand((char *)(strp + 1), imacro->im_lnbuf, LNSIZ); - } - - mjump_align = 1; - } +// 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); return imacro->im_lnbuf; } @@ -558,19 +638,17 @@ char * getmln(void) // // Get Next Line of Text from a Repeat Block // -char * getrln(void) +char * GetNextRepeatLine(void) { - IREPT * irept; - LONG * strp; - irept = cur_inobj->inobj.irept; - strp = irept->ir_nextln; // initial null + IREPT * irept = cur_inobj->inobj.irept; + LONG * strp = irept->ir_nextln; // initial null // Do repeat at end of .rept block's string list if (strp == NULL) { DEBUG printf("back-to-top-of-repeat-block count=%d\n", (int)irept->ir_count); - irept->ir_nextln = irept->ir_firstln; // copy first line + irept->ir_nextln = irept->ir_firstln; // copy first line if (irept->ir_count-- == 0) { @@ -578,11 +656,10 @@ char * getrln(void) return NULL; } - strp = irept->ir_nextln; //strp + strp = irept->ir_nextln; } - strcpy(irbuf, (char*)(irept->ir_nextln + 1)); - + strcpy(irbuf, (char *)(irept->ir_nextln + 1)); DEBUG printf("repeat line='%s'\n", irbuf); irept->ir_nextln = (LONG *)*strp; @@ -599,132 +676,88 @@ int include(int handle, char * fname) INOBJ * inobj; FILEREC * fr; + // Verbose mode if (verb_flag) - printf("[Including: %s]\n", fname); // Verbose mode + printf("[include: %s, cfileno=%u]\n", fname, cfileno); // Alloc and initialize include-descriptors inobj = a_inobj(SRC_IFILE); ifile = inobj->inobj.ifile; - ifile->ifhandle = handle; // Setup file handle - ifile->ifind = ifile->ifcnt = 0; // Setup buffer indices - ifile->ifoldlineno = curlineno; // Save old line number - ifile->ifoldfname = curfname; // Save old filename - ifile->ifno = cfileno; // Save old file number - cfileno = ++filecount; // Compute new file number - curfname = nstring(fname); // Set current filename (alloc storage) - curlineno = 0; // Start on line zero + ifile->ifhandle = handle; // Setup file handle + ifile->ifind = ifile->ifcnt = 0; // Setup buffer indices + ifile->ifoldlineno = curlineno; // Save old line number + 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 *)amem((LONG)sizeof(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); return OK; } -// -// Initialize Tokenizer -// -void init_token(void) -{ - int i; // Iterator - char * htab = "0123456789abcdefABCDEF"; // Hex character table - - lnsave = 0; // Don't save lines - curfname = ""; // No file, empty filename - filecount = (WORD)-1; - cfileno = (WORD)-1; // cfileno gets bumped to 0 - curlineno = 0; - totlines = 0; - etok = tokbuf; - f_inobj = NULL; - f_ifile = NULL; - f_imacro = NULL; - cur_inobj = NULL; - filerec = NULL; - last_fr = NULL; - lntag = SPACE; - - // Initialize hex, "dot" and tolower tables - for(i=0; i<128; ++i) - { - hextab[i] = -1; - dotxtab[i] = 0; - tolowertab[i] = (char)i; - } - - for(i=0; htab[i]!=EOS; ++i) - hextab[htab[i]] = (char)((i < 16) ? i : i - 6); - - for(i='A'; i<='Z'; ++i) - tolowertab[i] |= 0x20; - - // These characters are legal immediately after a period - dotxtab['b'] = DOTB; // .b .B .s .S - dotxtab['B'] = 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; // .l .L - dotxtab['I'] = DOTI; -} - - // // Pop the Current Input Level // int fpop(void) { - INOBJ * inobj; IFILE * ifile; IMACRO * imacro; LONG * p, * p1; - - inobj = cur_inobj; + INOBJ * inobj = cur_inobj; if (inobj != NULL) { - // Pop IFENT levels until we reach the conditional assembly context we were at when the - // input object was entered. + // Pop IFENT levels until we reach the conditional assembly context we + // were at when the input object was entered. while (ifent != inobj->in_ifent) - d_endif (); + d_endif(); - tok = inobj->in_otok; // Restore tok and otok + tok = inobj->in_otok; // Restore tok and otok etok = inobj->in_etok; switch (inobj->in_type) { - case SRC_IFILE: // Pop and release an IFILE + case SRC_IFILE: // Pop and release an IFILE if (verb_flag) printf("[Leaving: %s]\n", curfname); ifile = inobj->inobj.ifile; ifile->if_link = f_ifile; f_ifile = ifile; - close(ifile->ifhandle); // Close source file - curfname = ifile->ifoldfname; // Set current filename - curlineno = ifile->ifoldlineno; // Set current line# + close(ifile->ifhandle); // Close source file +if (verb_flag) 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# DEBUG printf("cfileno=%d ifile->ifno=%d\n", (int)cfileno, (int)ifile->ifno); - cfileno = ifile->ifno; // Restore current file number + cfileno = ifile->ifno; // Restore current file number +if (verb_flag) printf("[fpop: (post) cfileno=%d ifile->ifno=%d]\n", (int)cfileno, (int)ifile->ifno); break; - case SRC_IMACRO: // Pop and release an IMACRO + case SRC_IMACRO: // Pop and release an IMACRO imacro = inobj->inobj.imacro; imacro->im_link = f_imacro; f_imacro = imacro; break; - case SRC_IREPT: // Pop and release an IREPT + case SRC_IREPT: // Pop and release an IREPT DEBUG printf("dealloc IREPT\n"); p = inobj->inobj.irept->ir_firstln; @@ -750,39 +783,37 @@ int fpop(void) // Get line from file into buf, return NULL on EOF or ptr to the start of a // null-term line // -char * getln(void) +char * GetNextLine(void) { - IFILE * fl; int i, j; char * p, * d; - int readamt; - - readamt = -1; // 0 if last read() yeilded 0 bytes - fl = cur_inobj->inobj.ifile; + int readamt = -1; // 0 if last read() yeilded 0 bytes + IFILE * fl = cur_inobj->inobj.ifile; for(;;) { - // Scan for next end-of-line; handle stupid text formats by treating \r\n the same as \n. - // (lone '\r' at end of buffer means we have to check for '\n'). + // Scan for next end-of-line; handle stupid text formats by treating + // \r\n the same as \n. (lone '\r' at end of buffer means we have to + // check for '\n'). i = 0; j = fl->ifcnt; d = &fl->ifbuf[fl->ifind]; - for(p=d; i= j) { - break; // Look for '\n' to eat + break; // Look for '\n' to eat } else if (p[1] == '\n') { - ++i; + i++; } } @@ -794,7 +825,8 @@ char * getln(void) } } - // Handle hanging lines by ignoring them (Input file is exhausted, no \r or \n on last line) + // Handle hanging lines by ignoring them (Input file is exhausted, no + // \r or \n on last line) if (!readamt && fl->ifcnt) { fl->ifcnt = 0; @@ -810,14 +842,15 @@ char * getln(void) return &fl->ifbuf[fl->ifind]; } - // Relocate what's left of a line to the beginning of the buffer, and read some more of the - // file in; return NULL if the buffer's empty and on EOF. + // Relocate what's left of a line to the beginning of the buffer, and + // read some more of the file in; return NULL if the buffer's empty and + // on EOF. if (fl->ifind != 0) { p = &fl->ifbuf[fl->ifind]; d = &fl->ifbuf[fl->ifcnt & 1]; - for(i = 0; i < fl->ifcnt; ++i) + for(i=0; iifcnt; i++) *d++ = *p++; fl->ifind = fl->ifcnt & 1; @@ -835,25 +868,27 @@ char * getln(void) // // Tokenize a Line // -int tokln(void) +int TokenizeLine(void) { - char * ln = NULL; // Ptr to current position in line - char * p; // Random character ptr - TOKEN *tk; // Token-deposit ptr - int state = 0; // State for keyword detector - int j = 0; // Var for keyword detector - char c; // Random char - VALUE v; // Random value - char * nullspot = NULL; // Spot to clobber for SYMBOL terminatn - int stuffnull; // 1:terminate SYMBOL '\0' at *nullspot + char * ln = NULL; // Ptr to current position in line + char * p; // Random character ptr + TOKEN * tk; // Token-deposit ptr + int state = 0; // State for keyword detector + int j = 0; // Var for keyword detector + char c; // Random char + VALUE v; // Random value + char * nullspot = NULL; // Spot to clobber for SYMBOL terminatn + int stuffnull; // 1:terminate SYMBOL '\0' at *nullspot char c1; + int stringNum = 0; // Pointer to string locations in tokenized line - retry: +retry: - if (cur_inobj == NULL) // Return EOF if input stack is empty + if (cur_inobj == NULL) // Return EOF if input stack is empty return TKEOF; - // Get another line of input from the current input source: a file, a macro, or a repeat-block + // Get another line of input from the current input source: a file, + // a macro, or a repeat-block switch (cur_inobj->in_type) { // Include-file: @@ -862,24 +897,26 @@ int tokln(void) // o tag the listing-line with a space; // o kludge lines generated by Alcyon C. case SRC_IFILE: - if ((ln = getln()) == NULL) + if ((ln = GetNextLine()) == NULL) { - fpop(); // Pop input level - goto retry; // Try for more lines +if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IFILE...\n"); + fpop(); // Pop input level + goto retry; // Try for more lines } - ++curlineno; // Bump line number + curlineno++; // Bump line number lntag = SPACE; if (as68_flag) { - // AS68 compatibility, throw away all lines starting with back-quotes, tildes, or '*' + // AS68 compatibility, throw away all lines starting with + // back-quotes, tildes, or '*' // On other lines, turn the first '*' into a semi-colon. if (*ln == '`' || *ln == '~' || *ln == '*') *ln = ';'; else { - for(p=ln; *p!=EOS; ++p) + for(p=ln; *p!=EOS; p++) { if (*p == '*') { @@ -895,10 +932,10 @@ int tokln(void) // o Handle end-of-macro; // o tag the listing-line with an at (@) sign. case SRC_IMACRO: - if ((ln = getmln()) == NULL) + if ((ln = GetNextMacroLine()) == NULL) { - exitmac(); // Exit macro (pop args, do fpop(), etc) - goto retry; // Try for more lines... + ExitMacro(); // Exit macro (pop args, do fpop(), etc) + goto retry; // Try for more lines... } lntag = '@'; @@ -907,8 +944,9 @@ int tokln(void) // o Handle end-of-repeat-block; // o tag the listing-line with a pound (#) sign. case SRC_IREPT: - if ((ln = getrln()) == NULL) + if ((ln = GetNextRepeatLine()) == NULL) { +if (verb_flag) printf("TokenizeLine: Calling fpop() from SRC_IREPT...\n"); fpop(); goto retry; } @@ -917,19 +955,21 @@ int tokln(void) break; } - // Save text of the line. We only do this during listings and within macro-type blocks, - // since it is expensive to unconditionally copy every line. + // Save text of the line. We only do this during listings and within + // macro-type blocks, since it is expensive to unconditionally copy every + // line. if (lnsave) strcpy(lnbuf, ln); // General house-keeping - tok = tokeol; // Set "tok" to EOL in case of error - tk = etok; // Reset token ptr - stuffnull = 0; // Don't stuff nulls - ++totlines; // Bump total #lines assembled - - // See if the entire line is a comment. This is a win if the programmer puts in lots of comments - if (*ln == '*' || *ln == ';' || ((*ln == '/') && (*(ln+1) == '/'))) + tok = tokeol; // Set "tok" to EOL in case of error + tk = etok; // Reset token ptr + stuffnull = 0; // Don't stuff nulls + totlines++; // Bump total #lines assembled + + // See if the entire line is a comment. This is a win if the programmer + // puts in lots of comments + if (*ln == '*' || *ln == ';' || ((*ln == '/') && (*(ln + 1) == '/'))) goto goteol; // Main tokenization loop; @@ -942,39 +982,41 @@ int tokln(void) { // Skip whitespace, handle EOL while ((int)chrtab[*ln] & WHITE) - ++ln; + 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 termination is - // always one symbol behind, since there may be no place for a null in the case that - // an operator immediately follows the name. + // Handle start of symbol. Symbols are null-terminated in place. The + // termination is always one symbol behind, since there may be no place + // for a null in the case that an operator immediately follows the name. c = chrtab[*ln]; if (c & STSYM) { - if (stuffnull) // Terminate old symbol + if (stuffnull) // Terminate old symbol from previous pass *nullspot = EOS; - v = 0; // Assume no DOT attrib follows symbol + v = 0; // Assume no DOT attrib follows symbol stuffnull = 1; - p = nullspot = ln++; // Nullspot -> start of this symbol + p = nullspot = ln++; // Nullspot -> start of this symbol // Find end of symbol (and compute its length) - for(j=1; (int)chrtab[*ln]&CTSYM; ++j) - ++ln; + for(j=1; (int)chrtab[*ln]&CTSYM; j++) + ln++; - // Handle "DOT" special forms (like ".b") that follow a normal symbol or keyword: + // Handle "DOT" special forms (like ".b") that follow a normal + // symbol or keyword: if (*ln == '.') { - *ln++ = EOS; // Terminate symbol - stuffnull = 0; // And never try it again + *ln++ = EOS; // Terminate symbol + 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....) + // 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"); @@ -984,7 +1026,8 @@ int tokln(void) return error("misuse of `.', not allowed in symbols"); } - // If the symbol is small, check to see if it's really the name of a register. + // If the symbol is small, check to see if it's really the name of + // a register. if (j <= KWSIZE) { for(state=0; state>=0;) @@ -1012,6 +1055,12 @@ int tokln(void) j = -1; } + // Make j = -1 if user tries to use a RISC register while in 68K mode + if (!(rgpu || rdsp) && ((TOKEN)j >= KW_R0 && (TOKEN)j <= KW_R31)) + { + j = -1; + } + //make j = -1 if time, date etc with no preceeding ^^ //defined, referenced, streq, macdef, date and time switch ((TOKEN)j) @@ -1023,13 +1072,23 @@ int tokln(void) case 120: // time case 121: // date j = -1; - break; +// break; } + // If not tokenized keyword OR token was not found if (j < 0 || state < 0) { *tk++ = SYMBOL; +//#warning +//problem here: nullspot is a char * but TOKEN is a uint32_t. On a 64-bit system, +//this will cause all kinds of mischief. +#if 0 *tk++ = (TOKEN)nullspot; +#else + string[stringNum] = nullspot; + *tk++ = stringNum; + stringNum++; +#endif } else { @@ -1037,10 +1096,10 @@ int tokln(void) stuffnull = 0; } - if (v) // Record attribute token (if any) + if (v) // Record attribute token (if any) *tk++ = (TOKEN)v; - if (stuffnull) // Arrange for string termination + if (stuffnull) // Arrange for string termination on next pass nullspot = ln; continue; @@ -1072,7 +1131,16 @@ int tokln(void) case '\"': // "string" c1 = ln[-1]; *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 + string[stringNum] = ln; + *tk++ = stringNum; + stringNum++; +#endif for(p=ln; *ln!=EOS && *ln!=c1;) { @@ -1113,7 +1181,7 @@ int tokln(void) break; default: warn("bad backslash code in string"); - --ln; + ln--; break; } } @@ -1136,19 +1204,19 @@ int tokln(void) if (*ln == '.') { - if ((*(ln+1) == 'b') || (*(ln+1) == 'B')) + if ((*(ln + 1) == 'b') || (*(ln + 1) == 'B')) { v &= 0x000000FF; ln += 2; } - if ((*(ln+1) == 'w') || (*(ln+1) == 'W')) + if ((*(ln + 1) == 'w') || (*(ln + 1) == 'W')) { v &= 0x0000FFFF; ln += 2; } - if ((*(ln+1) == 'l') || (*(ln+1) == 'L')) + if ((*(ln + 1) == 'l') || (*(ln + 1) == 'L')) { ln += 2; } @@ -1205,11 +1273,11 @@ int tokln(void) { case '>': *tk++ = SHR; - ++ln; + ln++; continue; case '=': *tk++ = GE; - ++ln; + ln++; continue; default: *tk++ = '>'; @@ -1229,19 +1297,19 @@ int tokln(void) if (*ln == '.') { - if ((*(ln+1) == 'b') || (*(ln+1) == 'B')) + if ((*(ln + 1) == 'b') || (*(ln + 1) == 'B')) { v &= 0x000000FF; ln += 2; } - if ((*(ln+1) == 'w') || (*(ln+1) == 'W')) + if ((*(ln + 1) == 'w') || (*(ln + 1) == 'W')) { v &= 0x0000FFFF; ln += 2; } - if ((*(ln+1) == 'l') || (*(ln+1) == 'L')) + if ((*(ln + 1) == 'l') || (*(ln + 1) == 'L')) { ln += 2; } @@ -1350,21 +1418,20 @@ int tokln(void) while ((int)chrtab[*ln] & DIGIT) v = (v * 10) + *ln++ - '0'; + // See if there's a .[bwl] after the constant & deal with it if so if (*ln == '.') { - if ((*(ln+1) == 'b') || (*(ln+1) == 'B')) + if ((*(ln + 1) == 'b') || (*(ln + 1) == 'B')) { v &= 0x000000FF; ln += 2; } - - if ((*(ln+1) == 'w') || (*(ln+1) == 'W')) + else if ((*(ln + 1) == 'w') || (*(ln + 1) == 'W')) { v &= 0x0000FFFF; ln += 2; } - - if ((*(ln+1) == 'l') || (*(ln+1) == 'L')) + else if ((*(ln + 1) == 'l') || (*(ln + 1) == 'L')) { ln += 2; } @@ -1372,6 +1439,7 @@ int tokln(void) *tk++ = CONST; *tk++ = v; +//printf("CONST: %i\n", v); continue; } @@ -1382,9 +1450,9 @@ int tokln(void) // Terminate line of tokens and return "success." goteol: - tok = etok; // Set tok to beginning of line + tok = etok; // Set tok to beginning of line - if (stuffnull) // Terminate last SYMBOL + if (stuffnull) // Terminate last SYMBOL *nullspot = EOS; *tk++ = EOL; @@ -1409,35 +1477,44 @@ goteol: // be EOL. // //int d_goto(WORD siz) { -int d_goto(void) +//int d_goto(void) +int d_goto(WORD unused) { - char * sym; // Label to search for - LONG * defln; // Macro definition strings +// char * sym; // Label to search for +// LONG * defln; // Macro definition strings char * s1; // Temps for string comparison char * s2; - IMACRO * imacro; // Macro invocation block +// IMACRO * imacro; // Macro invocation block // Setup for the search if (*tok != SYMBOL) return error("missing label"); - sym = (char *)tok[1]; +// sym = (char *)tok[1]; + char * sym = string[tok[1]]; tok += 2; if (cur_inobj->in_type != SRC_IMACRO) return error("goto not in macro"); - imacro = cur_inobj->inobj.imacro; - defln = (LONG *)imacro->im_macro->svalue; + IMACRO * imacro = cur_inobj->inobj.imacro; +// defln = (LONG *)imacro->im_macro->svalue; + struct LineList * defln = imacro->im_macro->lineList; // Find the label, starting with the first line. - for(; defln!=NULL; defln=(LONG *)*defln) +// for(; defln!=NULL; defln=(LONG *)*defln) + for(; defln!=NULL; defln=defln->next) { - if (*(char *)(defln + 1) == ':') +// if (*(char *)(defln + 1) == ':') + if (defln->line[0] == ':') { // Compare names (sleazo string compare) + // This string compare is not right. Doesn't check for lengths. + // (actually it does, but in a crappy, unclear way.) +#warning "!!! Bad string comparison !!!" s1 = sym; - s2 = (char *)(defln + 1) + 1; +// s2 = (char *)(defln + 1) + 1; + s2 = defln->line; while (*s1 == *s2) { @@ -1445,8 +1522,8 @@ int d_goto(void) break; else { - ++s1; - ++s2; + s1++; + s2++; } } @@ -1461,3 +1538,94 @@ int d_goto(void) return error("goto label not found"); } + + +void DumpTokenBuffer(void) +{ + TOKEN * t; + printf("Tokens [%X]: ", sloc); + + for(t=tokbuf; *t!=EOL; t++) + { + if (*t == COLON) + printf("[COLON]"); + else if (*t == CONST) + { + t++; + printf("[CONST: $%X]", (uint32_t)*t); + } + else if (*t == ACONST) + printf("[ACONST]"); + else if (*t == STRING) +// printf("[STRING]"); + { + t++; + printf("[STRING:\"%s\"]", string[*t]); + } + else if (*t == SYMBOL) + { + t++; + printf("[SYMBOL:\"%s\"]", string[*t]); + } + else if (*t == EOS) + printf("[EOS]"); + else if (*t == TKEOF) + printf("[TKEOF]"); + else if (*t == DEQUALS) + printf("[DEQUALS]"); + else if (*t == SET) + printf("[SET]"); + else if (*t == REG) + printf("[REG]"); + else if (*t == DCOLON) + printf("[DCOLON]"); + else if (*t == GE) + printf("[GE]"); + else if (*t == LE) + printf("[LE]"); + else if (*t == NE) + printf("[NE]"); + else if (*t == SHR) + printf("[SHR]"); + else if (*t == SHL) + printf("[SHL]"); + else if (*t == UNMINUS) + printf("[UNMINUS]"); + else if (*t == DOTB) + printf("[DOTB]"); + else if (*t == DOTW) + printf("[DOTW]"); + else if (*t == DOTL) + printf("[DOTL]"); + else if (*t == DOTI) + printf("[DOTI]"); + else if (*t == ENDEXPR) + printf("[ENDEXPR]"); + else if (*t == CR_DEFINED) + printf("[CR_DEFINED]"); + else if (*t == CR_REFERENCED) + printf("[CR_REFERENCED]"); + else if (*t == CR_STREQ) + printf("[CR_STREQ]"); + else if (*t == CR_MACDEF) + printf("[CR_MACDEF]"); + else if (*t == CR_TIME) + printf("[CR_TIME]"); + else if (*t == CR_DATE) + printf("[CR_DATE]"); + else if (*t >= 0x20 && *t <= 0x2F) + printf("[%c]", (char)*t); + else if (*t >= 0x3A && *t <= 0x3F) + printf("[%c]", (char)*t); + else if (*t >= 0x80 && *t <= 0x87) + printf("[D%u]", ((uint32_t)*t) - 0x80); + else if (*t >= 0x88 && *t <= 0x8F) + printf("[A%u]", ((uint32_t)*t) - 0x88); + else + printf("[%X:%c]", (uint32_t)*t, (char)*t); +// printf("[%X]", (uint32_t)*t); + } + + printf("[EOL]\n"); +} +