From 210aef28ac574cefab31a79da0e9a050f1ed8bb9 Mon Sep 17 00:00:00 2001
From: Lu Wang <coolwanglu@gmail.com>
Date: Tue, 7 Aug 2012 01:52:31 +0800
Subject: [PATCH] better scripting; working on encoding

---
 bin/convert.pe      |  1 +
 bin/pdf2htmlEX      | 17 ++++++++++++++---
 src/HTMLRenderer.cc | 43 ++++++++-----------------------------------
 src/HTMLRenderer.h  |  6 ++----
 4 files changed, 25 insertions(+), 42 deletions(-)

diff --git a/bin/convert.pe b/bin/convert.pe
index 6c195bf..aa0c124 100644
--- a/bin/convert.pe
+++ b/bin/convert.pe
@@ -1,2 +1,3 @@
 Open($1);
+Reencode("unicode");
 Generate($1:r+".ttf");
diff --git a/bin/pdf2htmlEX b/bin/pdf2htmlEX
index 7261b0e..38d66f6 100755
--- a/bin/pdf2htmlEX
+++ b/bin/pdf2htmlEX
@@ -6,11 +6,22 @@ SCRIPT_DIR="$( cd -P "$( dirname "$SOURCE" )" && pwd )"/
 # Execute
 ${SCRIPT_DIR}/pdftohtmlEX $* 
 
+echo -n "Converting fonts: "
+
 for f in *.ttf; do
-    fontforge -script "${SCRIPT_DIR}/convert.pe" $f 2>/dev/null
+    if [ -f $f ]; then
+        fontforge -script "${SCRIPT_DIR}/convert.pe" $f 2>/dev/null
+        echo -n "."
+    fi
 done
 
 for f in *.pfa; do
-    fontforge -script "${SCRIPT_DIR}/convert.pe" $f 2>/dev/null
-#    rm $f
+    if [ -f $f ]; then
+        fontforge -script "${SCRIPT_DIR}/convert.pe" $f 2>/dev/null
+        rm $f
+        echo -n "."
+    fi
 done
+
+echo
+echo "Done."
diff --git a/src/HTMLRenderer.cc b/src/HTMLRenderer.cc
index 67c53be..991ebbc 100644
--- a/src/HTMLRenderer.cc
+++ b/src/HTMLRenderer.cc
@@ -143,20 +143,16 @@ void TextString::addChars(GfxState *state, double x, double y,
     height += dy;
 }
 
-void TextString::addChar(GfxState *state, double x, double y,
-        double dx, double dy, Unicode u)
+void TextString::addUnicodes(GfxState *state, double x, double y,
+        double dx, double dy, Unicode * u, int uLen)
 {
+    /*
     if (0 < u && u != 9 && u < 32)	// skip non-printable not-tab character
         return;
+        */
 
-    /*
-    if (unicodes.empty())
-    {
-        this->x = x;
-        this->y = y;
-    }
-    */
-    unicodes.push_back(u);
+    for(int i = 0; i < uLen; ++i)
+        unicodes.push_back(u[i]);
 
     width += dx;
     height += dy;
@@ -226,7 +222,6 @@ void HTMLRenderer::process(PDFDoc *doc)
     delete bg_renderer;
 
     std::cerr << std::endl;
-    std::cerr << "Done." << std::endl;
 }
 
 void HTMLRenderer::startPage(int pageNum, GfxState *state) 
@@ -459,29 +454,12 @@ void HTMLRenderer::drawChar(GfxState *state, double x, double y,
         double originX, double originY,
         CharCode code, int nBytes, Unicode *u, int uLen)
 {
-    double x1, y1, w1, h1;
-    
-    x1 = x;
-    y1 = y;
-
     // if it is hidden, then return
     if ((state->getRender() & 3) == 3)
         return ;
 
-    w1 = dx - state->getCharSpace() * state->getHorizScaling(),
-    h1 = dy;
-
-    cur_string->addChars(state, x1, y1, w1, h1, code, nBytes);
-
-    /*
-    if (uLen != 0) {
-        w1 /= uLen;
-        h1 /= uLen;
-    }
-    for (int i = 0; i < uLen; ++i) {
-        cur_string->addChar(state, x1 + i*w1, y1 + i*h1, w1, h1, u[i]);
-    }
-    */
+    //cur_string->addChars(state, x, y, dx, dy, code, nBytes);
+    cur_string->addUnicodes(state, x, y, dx, dy, u, uLen);
 }
 
 // TODO
@@ -546,11 +524,6 @@ long long HTMLRenderer::install_font(GfxFont * font)
         return new_fn_id;
     }
 
-    //debug
-    {
-        std::cerr << "install font: " << new_fn_id << ' ' << font->getID()->num << std::endl;
-    }
-
     string new_fn = (boost::format("f%|1$x|") % new_fn_id).str();
 
     if(font->getType() == fontType3) {
diff --git a/src/HTMLRenderer.h b/src/HTMLRenderer.h
index f39aedc..dd1fe54 100644
--- a/src/HTMLRenderer.h
+++ b/src/HTMLRenderer.h
@@ -52,11 +52,9 @@ class TextString
         void addChars(GfxState * state, double x, double y,
                 double dx, double dy,
                 CharCode code, int nbytes);
-
-        // Add a character to the string.
-        void addChar(GfxState *state, double x, double y,
+        void addUnicodes(GfxState *state, double x, double y,
                 double dx, double dy,
-                Unicode u);
+                Unicode *u, int uLen);
         double getX() const {return x;}
         double getY() const {return y;}
         double getWidth() const {return width;}