#18 hotfix patch for bug #845283 (\fe\ff in titles)

[Andreas Baumann]

...see bugreport.

backport of code from xpdf 3.00 src/pdfinfo.cc.
The code is NOT checked for memory leaks as the
code has big problems with memory managament
anyway.

[Andreas Baumann]

Logged In: YES
user_id=574221

diff -ruN pdftohtml-0.36/src/pdftohtml.cc
pdftohtml-0.36-bugfix#845283/src/pdftohtml.cc
--- pdftohtml-0.36/src/pdftohtml.cc 2003-06-25
00:41:27.000000000 +0200
+++ pdftohtml-0.36-bugfix#845283/src/pdftohtml.cc
2004-10-07 16:14:54.000000000 +0200
@@ -49,11 +49,12 @@
GBool noMerge = gFalse;
static char ownerPassword[33] = "";
static char userPassword[33] = "";
+static char cfgFileName[256] = "";
static char gsDevice[33] = "png16m";
static GBool printVersion = gFalse;

-static GString* getInfoString(Dict *infoDict, char *key);
-static GString* getInfoDate(Dict *infoDict, char *key);
+static GString* getInfoString(Dict *infoDict, char *key,
UnicodeMap *uMap);
+static GString* getInfoDate(Dict *infoDict, char *key,
UnicodeMap *uMap);

static char textEncName[128] = "";

@@ -98,6 +99,8 @@
"owner password (for encrypted files)"},
{"-upw", argString, userPassword,
sizeof(userPassword),
"user password (for encrypted files)"},
+ {"-cfg", argString, cfgFileName,
sizeof(cfgFileName),
+ "configuration file to use in place of .xpdfrc"},
{NULL}
};

@@ -116,6 +119,7 @@
GString *ownerPW, *userPW;
Object info;
char * extsList[] = {"png", "jpeg", "bmp", "pcx", "tiff",
"pbm", NULL};
+ UnicodeMap *uMap;

// parse args
ok = parseArgs(argDesc, &argc, argv);
@@ -147,6 +151,18 @@
}
}

+ // read config file
+ globalParams = new GlobalParams(cfgFileName);
+ if (textEncName[0]) {
+ globalParams->setTextEncoding(textEncName);
+ }
+
+ // get mapping to output encoding
+ if (!(uMap = globalParams->getTextEncoding())) {
+ error(-1, "Couldn't get text encoding");
+ goto error;
+ }
+
// open PDF file
if (ownerPassword[0]) {
ownerPW = new GString(ownerPassword);
@@ -232,13 +248,13 @@

doc->getDocInfo(&info);
if (info.isDict()) {
- docTitle = getInfoString(info.getDict(), "Title");
- author = getInfoString(info.getDict(), "Author");
- keywords = getInfoString(info.getDict(), "Keywords");
- subject = getInfoString(info.getDict(), "Subject");
- date = getInfoDate(info.getDict(), "ModDate");
+ docTitle = getInfoString(info.getDict(), "Title", uMap);
+ author = getInfoString(info.getDict(), "Author", uMap);
+ keywords = getInfoString(info.getDict(), "Keywords",
uMap);
+ subject = getInfoString(info.getDict(), "Subject", uMap);
+ date = getInfoDate(info.getDict(), "ModDate", uMap);
if( !date )
- date = getInfoDate(info.getDict(), "CreationDate");
+ date = getInfoDate(info.getDict(), "CreationDate",
uMap);
}
info.free();
if( !docTitle ) docTitle = new GString(htmlFileName);
@@ -365,18 +381,44 @@
return 0;
}

-static GString* getInfoString(Dict *infoDict, char *key) {
+static GString* getInfoString(Dict *infoDict, char *key,
UnicodeMap *uMap) {
Object obj;
GString *s1 = NULL;
+ GString *s2 = NULL;
+ GBool isUnicode;
+ Unicode u;
+ char buf[8];
+ int i, n;

if (infoDict->lookup(key, &obj)->isString()) {
s1 = new GString(obj.getString());
+ s2 = new GString("");
+ if ((s1->getChar(0) & 0xff) == 0xfe &&
+ (s1->getChar(1) & 0xff) == 0xff) {
+ isUnicode = gTrue;
+ i = 2;
+ } else {
+ isUnicode = gFalse;
+ i = 0;
+ }
+ while (i < obj.getString()->getLength()) {
+ if (isUnicode) {
+ u = ((s1->getChar(i) & 0xff) << 8) |
+ (s1->getChar(i+1) & 0xff);
+ i += 2;
+ } else {
+ u = s1->getChar(i) & 0xff;
+ ++i;
+ }
+ n = uMap->mapUnicode(u, buf, sizeof(buf));
+ s2->append(buf, n);
+ }
}
obj.free();
- return s1;
+ return s2;
}

-static GString* getInfoDate(Dict *infoDict, char *key) {
+static GString* getInfoDate(Dict *infoDict, char *key,
UnicodeMap *uMap) {
Object obj;
char *s;
int year, mon, day, hour, min, sec;

[Lucas B.]

Logged In: YES
user_id=964337

this is not fixed in 0.39
this bug also applies to the titles of the elements in
*-outline.html. with similar changes in
HtmlOutputDev::newOutlineLevel and possibly some other
places you might fix this bug completely.