Diffstat (limited to 'noncore/unsupported/qpdf/xpdf/CharCodeToUnicode.cc') (more/less context) (ignore whitespace changes)
-rw-r--r-- | noncore/unsupported/qpdf/xpdf/CharCodeToUnicode.cc | 160 |
1 files changed, 78 insertions, 82 deletions
diff --git a/noncore/unsupported/qpdf/xpdf/CharCodeToUnicode.cc b/noncore/unsupported/qpdf/xpdf/CharCodeToUnicode.cc index 6793398..912981e 100644 --- a/noncore/unsupported/qpdf/xpdf/CharCodeToUnicode.cc +++ b/noncore/unsupported/qpdf/xpdf/CharCodeToUnicode.cc @@ -1,288 +1,284 @@ //======================================================================== // // CharCodeToUnicode.cc // -// Copyright 2001 Derek B. Noonburg +// Copyright 2001-2002 Glyph & Cog, LLC // //======================================================================== #ifdef __GNUC__ #pragma implementation #endif #include <aconf.h> #include <stdio.h> #include <string.h> #include "gmem.h" #include "gfile.h" #include "GString.h" #include "Error.h" #include "GlobalParams.h" +#include "PSTokenizer.h" #include "CharCodeToUnicode.h" //------------------------------------------------------------------------ #define maxUnicodeString 8 struct CharCodeToUnicodeString { CharCode c; Unicode u[maxUnicodeString]; int len; }; //------------------------------------------------------------------------ +static int getCharFromString(void *data) { + char *p; + int c; + + p = *(char **)data; + if (*p) { + c = *p++; + *(char **)data = p; + } else { + c = EOF; + } + return c; +} + +static int getCharFromFile(void *data) { + return fgetc((FILE *)data); +} + +//------------------------------------------------------------------------ + CharCodeToUnicode *CharCodeToUnicode::parseCIDToUnicode(GString *collectionA) { FILE *f; Unicode *mapA; CharCode size, mapLenA; char buf[64]; Unicode u; CharCodeToUnicode *ctu; if (!(f = globalParams->getCIDToUnicodeFile(collectionA))) { error(-1, "Couldn't find cidToUnicode file for the '%s' collection", collectionA->getCString()); return NULL; } size = 32768; mapA = (Unicode *)gmalloc(size * sizeof(Unicode)); mapLenA = 0; while (getLine(buf, sizeof(buf), f)) { if (mapLenA == size) { size *= 2; mapA = (Unicode *)grealloc(mapA, size * sizeof(Unicode)); } if (sscanf(buf, "%x", &u) == 1) { mapA[mapLenA] = u; } else { error(-1, "Bad line (%d) in cidToUnicode file for the '%s' collection", (int)(mapLenA + 1), collectionA->getCString()); mapA[mapLenA] = 0; } ++mapLenA; } ctu = new CharCodeToUnicode(collectionA->copy(), mapA, mapLenA, gTrue, NULL, 0); gfree(mapA); return ctu; } CharCodeToUnicode *CharCodeToUnicode::make8BitToUnicode(Unicode *toUnicode) { return new CharCodeToUnicode(NULL, toUnicode, 256, gTrue, NULL, 0); } -static char *getLineFromString(char *buf, int size, char **s) { - char c; - int i; - - i = 0; - while (i < size - 1 && **s) { - buf[i++] = c = *(*s)++; - if (c == '\x0a') { - break; - } - if (c == '\x0d') { - if (**s == '\x0a' && i < size - 1) { - buf[i++] = '\x0a'; - ++*s; - } - break; - } - } - buf[i] = '\0'; - if (i == 0) { - return NULL; - } - return buf; -} - CharCodeToUnicode *CharCodeToUnicode::parseCMap(GString *buf, int nBits) { CharCodeToUnicode *ctu; char *p; ctu = new CharCodeToUnicode(NULL); p = buf->getCString(); - ctu->parseCMap1((char *(*)(char *, int, void *))&getLineFromString, - &p, nBits); + ctu->parseCMap1(&getCharFromString, &p, nBits); return ctu; } -void CharCodeToUnicode::parseCMap1(char *(*getLineFunc)(char *, int, void *), - void *data, int nBits) { - char buf[256]; - GBool inBFChar, inBFRange; - char *tok1, *tok2, *tok3; +void CharCodeToUnicode::parseCMap1(int (*getCharFunc)(void *), void *data, + int nBits) { + PSTokenizer *pst; + char tok1[256], tok2[256], tok3[256]; int nDigits, n1, n2, n3; CharCode oldLen, i; CharCode code1, code2; Unicode u; char uHex[5]; int j; GString *name; FILE *f; nDigits = nBits / 4; - inBFChar = inBFRange = gFalse; - while ((*getLineFunc)(buf, sizeof(buf), data)) { - tok1 = strtok(buf, " \t\r\n"); - if (!tok1 || tok1[0] == '%') { - continue; - } - tok2 = strtok(NULL, " \t\r\n"); - tok3 = strtok(NULL, " \t\r\n"); - if (inBFChar) { - if (!strcmp(tok1, "endbfchar")) { - inBFChar = gFalse; - } else if (tok2) { - n1 = strlen(tok1); - n2 = strlen(tok2); + pst = new PSTokenizer(getCharFunc, data); + pst->getToken(tok1, sizeof(tok1), &n1); + while (pst->getToken(tok2, sizeof(tok2), &n2)) { + if (!strcmp(tok2, "usecmap")) { + if (tok1[0] == '/') { + name = new GString(tok1 + 1); + if ((f = globalParams->findToUnicodeFile(name))) { + parseCMap1(&getCharFromFile, f, nBits); + fclose(f); + } else { + error(-1, "Couldn't find ToUnicode CMap file for '%s'", + name->getCString()); + } + delete name; + } + pst->getToken(tok1, sizeof(tok1), &n1); + } else if (!strcmp(tok2, "beginbfchar")) { + while (pst->getToken(tok1, sizeof(tok1), &n1)) { + if (!strcmp(tok1, "endbfchar")) { + break; + } + if (!pst->getToken(tok2, sizeof(tok2), &n2) || + !strcmp(tok2, "endbfchar")) { + error(-1, "Illegal entry in bfchar block in ToUnicode CMap"); + break; + } if (!(n1 == 2 + nDigits && tok1[0] == '<' && tok1[n1 - 1] == '>' && tok2[0] == '<' && tok2[n2 - 1] == '>')) { - error(-1, "Illegal line in bfchar block in ToUnicode CMap"); + error(-1, "Illegal entry in bfchar block in ToUnicode CMap"); continue; } tok1[n1 - 1] = tok2[n2 - 1] = '\0'; if (sscanf(tok1 + 1, "%x", &code1) != 1) { - error(-1, "Illegal line in bfchar block in ToUnicode CMap"); + error(-1, "Illegal entry in bfchar block in ToUnicode CMap"); continue; } if (code1 >= mapLen) { oldLen = mapLen; mapLen = (code1 + 256) & ~255; map = (Unicode *)grealloc(map, mapLen * sizeof(Unicode)); for (i = oldLen; i < mapLen; ++i) { map[i] = 0; } } if (n2 == 6) { if (sscanf(tok2 + 1, "%x", &u) != 1) { - error(-1, "Illegal line in bfchar block in ToUnicode CMap"); + error(-1, "Illegal entry in bfchar block in ToUnicode CMap"); continue; } map[code1] = u; } else { map[code1] = 0; if (sMapLen == sMapSize) { sMapSize += 8; sMap = (CharCodeToUnicodeString *) grealloc(sMap, sMapSize * sizeof(CharCodeToUnicodeString)); } sMap[sMapLen].c = code1; sMap[sMapLen].len = (n2 - 2) / 4; for (j = 0; j < sMap[sMapLen].len && j < maxUnicodeString; ++j) { strncpy(uHex, tok2 + 1 + j*4, 4); uHex[4] = '\0'; if (sscanf(uHex, "%x", &sMap[sMapLen].u[j]) != 1) { - error(-1, "Illegal line in bfchar block in ToUnicode CMap"); + error(-1, "Illegal entry in bfchar block in ToUnicode CMap"); } } ++sMapLen; } - } else { - error(-1, "Illegal bfchar block in ToUnicode CMap"); } - } else if (inBFRange) { - if (!strcmp(tok1, "endbfrange")) { - inBFRange = gFalse; - } else if (tok2 && tok3) { - n1 = strlen(tok1); - n2 = strlen(tok2); - n3 = strlen(tok3); + pst->getToken(tok1, sizeof(tok1), &n1); + } else if (!strcmp(tok2, "beginbfrange")) { + while (pst->getToken(tok1, sizeof(tok1), &n1)) { + if (!strcmp(tok1, "endbfrange")) { + break; + } + if (!pst->getToken(tok2, sizeof(tok2), &n2) || + !strcmp(tok2, "endbfrange") || + !pst->getToken(tok3, sizeof(tok3), &n3) || + !strcmp(tok3, "endbfrange")) { + error(-1, "Illegal entry in bfrange block in ToUnicode CMap"); + break; + } if (!(n1 == 2 + nDigits && tok1[0] == '<' && tok1[n1 - 1] == '>' && n2 == 2 + nDigits && tok2[0] == '<' && tok2[n2 - 1] == '>' && tok3[0] == '<' && tok3[n3 - 1] == '>')) { - error(-1, "Illegal line in bfrange block in ToUnicode CMap"); + error(-1, "Illegal entry in bfrange block in ToUnicode CMap"); continue; } tok1[n1 - 1] = tok2[n2 - 1] = tok3[n3 - 1] = '\0'; if (sscanf(tok1 + 1, "%x", &code1) != 1 || sscanf(tok2 + 1, "%x", &code2) != 1) { - error(-1, "Illegal line in bfrange block in ToUnicode CMap"); + error(-1, "Illegal entry in bfrange block in ToUnicode CMap"); continue; } if (code2 >= mapLen) { oldLen = mapLen; mapLen = (code2 + 256) & ~255; map = (Unicode *)grealloc(map, mapLen * sizeof(Unicode)); for (i = oldLen; i < mapLen; ++i) { map[i] = 0; } } if (n3 == 6) { if (sscanf(tok3 + 1, "%x", &u) != 1) { - error(-1, "Illegal line in bfrange block in ToUnicode CMap"); + error(-1, "Illegal entry in bfrange block in ToUnicode CMap"); continue; } for (; code1 <= code2; ++code1) { map[code1] = u++; } } else { if (sMapLen + (int)(code2 - code1 + 1) > sMapSize) { sMapSize = (sMapSize + (code2 - code1 + 1) + 7) & ~7; sMap = (CharCodeToUnicodeString *) grealloc(sMap, sMapSize * sizeof(CharCodeToUnicodeString)); } for (i = 0; code1 <= code2; ++code1, ++i) { map[code1] = 0; sMap[sMapLen].c = code1; sMap[sMapLen].len = (n3 - 2) / 4; for (j = 0; j < sMap[sMapLen].len && j < maxUnicodeString; ++j) { strncpy(uHex, tok3 + 1 + j*4, 4); uHex[4] = '\0'; if (sscanf(uHex, "%x", &sMap[sMapLen].u[j]) != 1) { - error(-1, "Illegal line in bfrange block in ToUnicode CMap"); + error(-1, "Illegal entry in bfrange block in ToUnicode CMap"); } } sMap[sMapLen].u[sMap[sMapLen].len - 1] += i; ++sMapLen; } } - } else { - error(-1, "Illegal bfrange block in ToUnicode CMap"); } - } else if (tok2 && !strcmp(tok2, "usecmap")) { - if (tok1[0] == '/') { - name = new GString(tok1 + 1); - if ((f = globalParams->findToUnicodeFile(name))) { - parseCMap1((char *(*)(char *, int, void *))&getLine, f, nBits); - fclose(f); - } else { - error(-1, "Couldn't find ToUnicode CMap file for '%s'", - name->getCString()); - } - delete name; - } - } else if (tok2 && !strcmp(tok2, "beginbfchar")) { - inBFChar = gTrue; - } else if (tok2 && !strcmp(tok2, "beginbfrange")) { - inBFRange = gTrue; + pst->getToken(tok1, sizeof(tok1), &n1); + } else { + strcpy(tok1, tok2); } } + delete pst; } CharCodeToUnicode::CharCodeToUnicode(GString *collectionA) { CharCode i; collection = collectionA; mapLen = 256; map = (Unicode *)gmalloc(mapLen * sizeof(Unicode)); for (i = 0; i < mapLen; ++i) { map[i] = 0; } sMap = NULL; sMapLen = sMapSize = 0; refCnt = 1; } CharCodeToUnicode::CharCodeToUnicode(GString *collectionA, Unicode *mapA, CharCode mapLenA, GBool copyMap, CharCodeToUnicodeString *sMapA, int sMapLenA) { collection = collectionA; mapLen = mapLenA; if (copyMap) { map = (Unicode *)gmalloc(mapLen * sizeof(Unicode)); |