-rw-r--r-- | library/backend/vcc.y | 103 |
1 files changed, 56 insertions, 47 deletions
diff --git a/library/backend/vcc.y b/library/backend/vcc.y index e326a64..5bcf0cb 100644 --- a/library/backend/vcc.y +++ b/library/backend/vcc.y @@ -424,193 +424,193 @@ static void enterValues(const char *value) } /* else this field is empty, advance to next field */ fieldedProp++; } else { if (value) { setVObjectStringZValue_(curProp,strdup( value )); } } deleteStr(value); } static void enterProps(const char *s) { curProp = addGroup(curObj,s); deleteStr(s); } static void enterAttr(const char *s1, const char *s2) { const char *p1, *p2; p1 = lookupProp_(s1); if (s2) { VObject *a; p2 = lookupProp_(s2); a = addProp(curProp,p1); setVObjectStringZValue(a,p2); } else addProp(curProp,p1); if (qstricmp(p1,VCBase64Prop) == 0 || (s2 && qstricmp(p2,VCBase64Prop)==0)) lexPushMode(L_BASE64); else if (qstricmp(p1,VCQuotedPrintableProp) == 0 || (s2 && qstricmp(p2,VCQuotedPrintableProp)==0)) lexPushMode(L_QUOTED_PRINTABLE); deleteStr(s1); deleteStr(s2); } #define MAX_LEX_LOOKAHEAD_0 32 #define MAX_LEX_LOOKAHEAD 64 #define MAX_LEX_MODE_STACK_SIZE 10 #define LEXMODE() (lexBuf.lexModeStack[lexBuf.lexModeStackTop]) struct LexBuf { /* input */ #ifdef INCLUDEMFC CFile *inputFile; #else FILE *inputFile; #endif char *inputString; unsigned long curPos; unsigned long inputLen; /* lookahead buffer */ /* -- lookahead buffer is short instead of char so that EOF / can be represented correctly. */ unsigned long len; short buf[MAX_LEX_LOOKAHEAD]; unsigned long getPtr; /* context stack */ unsigned long lexModeStackTop; enum LexMode lexModeStack[MAX_LEX_MODE_STACK_SIZE]; /* token buffer */ unsigned long maxToken; char *strs; unsigned long strsLen; } lexBuf; static void lexPushMode(enum LexMode mode) { if (lexBuf.lexModeStackTop == (MAX_LEX_MODE_STACK_SIZE-1)) yyerror("lexical context stack overflow"); else { lexBuf.lexModeStack[++lexBuf.lexModeStackTop] = mode; } } static void lexPopMode(int top) { /* special case of pop for ease of error recovery -- this version will never underflow */ if (top) lexBuf.lexModeStackTop = 0; else if (lexBuf.lexModeStackTop > 0) lexBuf.lexModeStackTop--; } static int lexWithinMode(enum LexMode mode) { unsigned long i; for (i=0;i<lexBuf.lexModeStackTop;i++) if (mode == lexBuf.lexModeStack[i]) return 1; return 0; } -static char lexGetc_() +static int lexGetc_()
{ /* get next char from input, no buffering. */ if (lexBuf.curPos == lexBuf.inputLen) return EOF; else if (lexBuf.inputString) return *(lexBuf.inputString + lexBuf.curPos++); else { #ifdef INCLUDEMFC char result; return lexBuf.inputFile->Read(&result, 1) == 1 ? result : EOF; #else return fgetc(lexBuf.inputFile); #endif } } static int lexGeta() { ++lexBuf.len; return (lexBuf.buf[lexBuf.getPtr] = lexGetc_()); } static int lexGeta_(int i) { ++lexBuf.len; return (lexBuf.buf[(lexBuf.getPtr+i)%MAX_LEX_LOOKAHEAD] = lexGetc_()); } static void lexSkipLookahead() { if (lexBuf.len > 0 && lexBuf.buf[lexBuf.getPtr]!=EOF) { /* don't skip EOF. */ lexBuf.getPtr = (lexBuf.getPtr + 1) % MAX_LEX_LOOKAHEAD; lexBuf.len--; } } static int lexLookahead() { int c = (lexBuf.len)? lexBuf.buf[lexBuf.getPtr]: lexGeta(); /* do the \r\n -> \n or \r -> \n translation here */ if (c == '\r') { int a = (lexBuf.len>1)? lexBuf.buf[(lexBuf.getPtr+1)%MAX_LEX_LOOKAHEAD]: lexGeta_(1); if (a == '\n') { lexSkipLookahead(); } lexBuf.buf[lexBuf.getPtr] = c = '\n'; } else if (c == '\n') { int a = (lexBuf.len>1)? lexBuf.buf[lexBuf.getPtr+1]: lexGeta_(1); if (a == '\r') { lexSkipLookahead(); } lexBuf.buf[lexBuf.getPtr] = '\n'; } return c; } static int lexGetc() { int c = lexLookahead(); if (lexBuf.len > 0 && lexBuf.buf[lexBuf.getPtr]!=EOF) { /* EOF will remain in lookahead buffer */ lexBuf.getPtr = (lexBuf.getPtr + 1) % MAX_LEX_LOOKAHEAD; lexBuf.len--; } return c; } static void lexSkipLookaheadWord() { if (lexBuf.strsLen <= lexBuf.len) { lexBuf.len -= lexBuf.strsLen; lexBuf.getPtr = (lexBuf.getPtr + lexBuf.strsLen) % MAX_LEX_LOOKAHEAD; } } static void lexClearToken() { lexBuf.strsLen = 0; } static void lexAppendc(int c) { lexBuf.strs[lexBuf.strsLen] = c; /* append up to zero termination */ if (c == 0) return; lexBuf.strsLen++; if (lexBuf.strsLen > lexBuf.maxToken) { /* double the token string size */ lexBuf.maxToken <<= 1; lexBuf.strs = (char*) realloc(lexBuf.strs,(size_t)lexBuf.maxToken); } } @@ -838,246 +838,255 @@ static char * lexGetDataFromBase64() else if (c == '+') b = 62; else if (c == '/') b = 63; else if (c == '=') { b = 0; pad++; } else if ((c == ' ') || (c == '\t')) { continue; } else { /* error condition */ if (bytes) free(bytes); else if (oldBytes) free(oldBytes); // error recovery: skip until 2 adjacent newlines. DBG_(("db: invalid character 0x%x '%c'\n", c,c)); if (c != EOF) { c = lexGetc(); while (c != EOF) { if (c == '\n' && lexLookahead() == '\n') { ++mime_lineNum; break; } c = lexGetc(); } } return NULL; } trip = (trip << 6) | b; if (++quadIx == 4) { unsigned char outBytes[3]; int numOut; int i; for (i = 0; i < 3; i++) { outBytes[2-i] = (unsigned char)(trip & 0xFF); trip >>= 8; } numOut = 3 - pad; if (bytesLen + numOut > bytesMax) { if (!bytes) { bytesMax = 1024; bytes = (unsigned char*)malloc((size_t)bytesMax); } else { bytesMax <<= 2; oldBytes = bytes; bytes = (unsigned char*)realloc(bytes,(size_t)bytesMax); } if (bytes == 0) { mime_error("out of memory while processing BASE64 data\n"); } } if (bytes) { memcpy(bytes + bytesLen, outBytes, numOut); bytesLen += numOut; } trip = 0; quadIx = 0; } } } /* while */ DBG_(("db: bytesLen = %d\n", bytesLen)); /* kludge: all this won't be necessary if we have tree form representation */ if (bytes) { setValueWithSize(curProp,bytes,(unsigned int)bytesLen); free(bytes); } else if (oldBytes) { setValueWithSize(curProp,oldBytes,(unsigned int)bytesLen); free(oldBytes); } return 0; } static int match_begin_end_name(int end) { int token; lexSkipWhite(); if (lexLookahead() != ':') return ID; lexSkipLookahead(); lexSkipWhite(); token = match_begin_name(end); if (token == ID) { lexPushLookaheadc(':'); DBG_(("db: ID '%s'\n", yylval.str)); return ID; } else if (token != 0) { lexSkipLookaheadWord(); deleteStr(yylval.str); DBG_(("db: begin/end %d\n", token)); return token; } return 0; } static char* lexGetQuotedPrintable() { - char cur; - + int c;
+ lexSkipWhite();
+ c = lexLookahead();
lexClearToken(); - do { - cur = lexGetc(); - switch (cur) { - case '=': { - int c = 0; - int next[2]; - int i; - for (i = 0; i < 2; i++) { - next[i] = lexGetc(); - if (next[i] >= '0' && next[i] <= '9') - c = c * 16 + next[i] - '0'; - else if (next[i] >= 'A' && next[i] <= 'F') - c = c * 16 + next[i] - 'A' + 10; - else +
+ while (c != EOF && c != ';') {
+ if (c == '\n') {
+ // break, leave '\n' on remaining chars.
break; + } else if (c == '=') {
+ int cur = 0;
+ int next;
+
+ lexSkipLookahead(); // skip '='
+ next = lexLookahead();
+
+ if (next == '\n') {
+ // skip and only skip the \n
+ lexSkipLookahead();
+ c = lexLookahead();
+ ++mime_lineNum; // aid in error reporting
+ continue;
+ } else if (next >= '0' && next <= '9') {
+ cur = next - '0';
+ } else if (next >= 'A' && next <= 'F') {
+ cur = next - 'A' + 10;
+ } else {
+ // we have been sent buggy stuff. doesn't matter
+ // what we do so long as we keep going.
+ // should probably spit an error here
+ c = lexLookahead();
+ continue;
} - if (i == 0) { - /* single '=' follow by LINESEP is continuation sign? */ - if (next[0] == '\n') { - ++mime_lineNum; - } - else { - lexPushLookaheadc('='); - goto EndString; - } +
+ lexSkipLookahead(); // skip A-Z0-9
+ next = lexLookahead();
+
+ cur = cur * 16;
+ // this time really just expecting 0-9A-F
+ if (next >= '0' && next <= '9') {
+ cur += next - '0';
+ } else if (next >= 'A' && next <= 'F') {
+ cur += next - 'A' + 10;
+ } else {
+ // we have been sent buggy stuff. doesn't matter
+ // what we do so long as we keep going.
+ // should probably spit an error here
+ c = lexLookahead();
+ continue;
} - else if (i == 1) { - lexPushLookaheadc(next[1]); - lexPushLookaheadc(next[0]); - lexAppendc('='); +
+ // got a valid escaped =. append it.
+ lexSkipLookahead(); // skip second 0-9A-F
+ lexAppendc(cur);
} else { - lexAppendc(c); + lexSkipLookahead(); // skip whatever we just read.
+ lexAppendc(c); // and append it.
} - break; - } /* '=' */ - case '\n': { - lexPushLookaheadc('\n'); - goto EndString; + c = lexLookahead();
} - case (char)EOF: - break; - default: - lexAppendc(cur); - break; - } /* switch */ - } while (cur != (char)EOF); - -EndString: lexAppendc(0); - return lexStr(); - } /* LexQuotedPrintable */ + return c==EOF?0:lexStr();
+}
static int yylex() { int lexmode = LEXMODE(); if (lexmode == L_VALUES) { int c = lexGetc(); if (c == ';') { DBG_(("db: SEMICOLON\n")); lexPushLookaheadc(c); handleMoreRFC822LineBreak(c); lexSkipLookahead(); return SEMICOLON; } else if (strchr("\n",c)) { ++mime_lineNum; /* consume all line separator(s) adjacent to each other */ c = lexLookahead(); while (strchr("\n",c)) { lexSkipLookahead(); c = lexLookahead(); ++mime_lineNum; } DBG_(("db: LINESEP\n")); return LINESEP; } else { char *p = 0; lexPushLookaheadc(c); if (lexWithinMode(L_BASE64)) { /* get each char and convert to bin on the fly... */ p = lexGetDataFromBase64(); yylval.str = p; return STRING; } else if (lexWithinMode(L_QUOTED_PRINTABLE)) { p = lexGetQuotedPrintable(); } else { #ifdef _SUPPORT_LINE_FOLDING p = lexGet1Value(); #else p = lexGetStrUntil(";\n"); #endif } if (p) { DBG_(("db: STRING: '%s'\n", p)); yylval.str = p; return STRING; } else return 0; } } else { /* normal mode */ while (1) { int c = lexGetc(); switch(c) { case ':': { /* consume all line separator(s) adjacent to each other */ /* ignoring linesep immediately after colon. */ c = lexLookahead(); while (strchr("\n",c)) { lexSkipLookahead(); c = lexLookahead(); ++mime_lineNum; } DBG_(("db: COLON\n")); return COLON; } case ';': DBG_(("db: SEMICOLON\n")); return SEMICOLON; case '=': DBG_(("db: EQ\n")); return EQ; /* ignore whitespace in this mode */ case '\t': case ' ': continue; case '\n': { ++mime_lineNum; continue; } case EOF: return 0; break; default: { lexPushLookaheadc(c); if (isalnum(c)) { char *t = lexGetWord(); yylval.str = t; if (!qstricmp(t, "begin")) { return match_begin_end_name(0); } else if (!qstricmp(t,"end")) { return match_begin_end_name(1); } else { |