Author: jahewson
Date: Fri Nov 6 00:01:55 2015
New Revision: 1712879
URL: http://svn.apache.org/viewvc?rev=1712879&view=rev
Log:
PDFBOX-2950: follow PDF spec rules for glyph mapping via CMap -> cmap
Modified:
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/cos/COSName.java
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdmodel/font/PDCIDFontType2.java
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdmodel/font/PDType0Font.java
Modified: pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/cos/COSName.java
URL:
http://svn.apache.org/viewvc/pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/cos/COSName.java?rev=1712879&r1=1712878&r2=1712879&view=diff
==============================================================================
--- pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/cos/COSName.java
(original)
+++ pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/cos/COSName.java Fri
Nov 6 00:01:55 2015
@@ -275,6 +275,7 @@ public final class COSName extends COSBa
public static final COSName ID_TREE = new COSName("IDTree");
public static final COSName IDENTITY = new COSName("Identity");
public static final COSName IDENTITY_H = new COSName("Identity-H");
+ public static final COSName IDENTITY_V = new COSName("Identity-V");
public static final COSName IF = new COSName("IF");
public static final COSName IM = new COSName("IM");
public static final COSName IMAGE = new COSName("Image");
Modified:
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdmodel/font/PDCIDFontType2.java
URL:
http://svn.apache.org/viewvc/pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdmodel/font/PDCIDFontType2.java?rev=1712879&r1=1712878&r2=1712879&view=diff
==============================================================================
---
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdmodel/font/PDCIDFontType2.java
(original)
+++
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdmodel/font/PDCIDFontType2.java
Fri Nov 6 00:01:55 2015
@@ -283,21 +283,13 @@ public class PDCIDFontType2 extends PDCI
// font's 'cmap' table. The means by which this is accomplished
are implementation-
// dependent.
- boolean hasUnicodeMap = parent.getCMapUCS2() != null;
-
if (cid2gid != null)
{
// Acrobat allows non-embedded GIDs - todo: can we find a test
PDF for this?
+ LOG.warn("Using non-embedded GIDs in font " + getName());
int cid = codeToCID(code);
return cid2gid[cid];
}
- else if (hasIdentityCid2Gid || !hasUnicodeMap)
- {
- // same as above, but for the default Identity CID2GIDMap or
when there is no
- // ToUnicode CMap to fallback to, see PDFBOX-2599 and
PDFBOX-2560
- // todo: can we find a test PDF for the Identity case?
- return codeToCID(code);
- }
else
{
// fallback to the ToUnicode CMap, test with PDFBOX-1422 and
PDFBOX-2560
Modified:
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdmodel/font/PDType0Font.java
URL:
http://svn.apache.org/viewvc/pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdmodel/font/PDType0Font.java?rev=1712879&r1=1712878&r2=1712879&view=diff
==============================================================================
---
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdmodel/font/PDType0Font.java
(original)
+++
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdmodel/font/PDType0Font.java
Fri Nov 6 00:01:55 2015
@@ -215,42 +215,35 @@ public class PDType0Font extends PDFont
*/
private void fetchCMapUCS2() throws IOException
{
- // if the font is composite and uses a predefined cmap (excluding
Identity-H/V) then
- // or if its decendant font uses Adobe-GB1/CNS1/Japan1/Korea1
- if (isCMapPredefined)
+ // if the font is composite and uses a predefined cmap (excluding
Identity-H/V)
+ // or whose descendant CIDFont uses the Adobe-GB1, Adobe-CNS1,
Adobe-Japan1, or
+ // Adobe-Korea1 character collection:
+ COSName name = dict.getCOSName(COSName.ENCODING);
+ if (isCMapPredefined && !(name == COSName.IDENTITY_H || name ==
COSName.IDENTITY_V) ||
+ isDescendantCJK)
{
// a) Map the character code to a CID using the font's CMap
// b) Obtain the ROS from the font's CIDSystemInfo
// c) Construct a second CMap name by concatenating the ROS in the
format "R-O-UCS2"
// d) Obtain the CMap with the constructed name
// e) Map the CID according to the CMap from step d), producing a
Unicode value
-
- String cMapName = null;
-
- // get the encoding CMap
- COSBase encoding = dict.getDictionaryObject(COSName.ENCODING);
- if (encoding instanceof COSName)
+
+ String strName = null;
+ if (name == null && isDescendantCJK)
{
- cMapName = ((COSName)encoding).getName();
+ strName = cMap.getRegistry() + "-" +
+ cMap.getOrdering() + "-" +
+ cMap.getSupplement();
}
-
- if ("Identity-H".equals(cMapName) || "Identity-V".equals(cMapName))
+ else if (name != null)
{
- if (isDescendantCJK)
- {
- cMapName = getCJKCMap(descendantFont.getCIDSystemInfo());
- }
- else
- {
- // we can't map Identity-H or Identity-V to Unicode
- return;
- }
+ strName = name.getName();
}
// try to find the corresponding Unicode (UC2) CMap
- if (cMapName != null)
+ if (strName != null)
{
- CMap cMap = CMapManager.getPredefinedCMap(cMapName);
+ CMap cMap = CMapManager.getPredefinedCMap(strName);
if (cMap != null)
{
String ucs2Name = cMap.getRegistry() + "-" +
cMap.getOrdering() + "-UCS2";
@@ -414,7 +407,7 @@ public class PDType0Font extends PDFont
return unicode;
}
- if (isCMapPredefined && cMapUCS2 != null)
+ if ((isCMapPredefined || isDescendantCJK) && cMapUCS2 != null)
{
// if the font is composite and uses a predefined cmap (excluding
Identity-H/V) then
// or if its decendant font uses Adobe-GB1/CNS1/Japan1/Korea1
