[FOP-1617] FOP gets an error for Windows Japanese font.

• Type: Bug
• Status: Resolved
• Resolution: Cannot Reproduce
• Affects Version/s: 0.95
• Component/s: font/unqualified
• Environment: Operating System: Windows XP; Platform: PC
• External issue ID: 46460
• Assignee: Unassigned
• Reporter: Mamoru Asagami
• Created: 01/Jan/09 00:00
• Updated: 16/May/18 12:54
• Resolved: 16/May/18 12:54

Description

Problem:

You get:
[Fatal Error] msmincho.xml:1:171: Character reference "&#0" is an invalid XML character

if you make font metrics files from msmincho.ttc and msgothic.ttc of Windows and use them.

To reproduce it, set up the attached file: fopdoc.zip as c:\fopdoc, fop.xconf.xml(one of the zip) as ${fop.home}/conf/fop.xconf.xml, and run ant as follows:

c:\fopdoc>ant

For your convenience, I included ant's error message as err.txt in the attached file: fopdoc.zip.

Solution:

I found something in "public final String readTTFString(int len)" of org.apache.fop.fonts.truetype.FontFileReader.

When I replaced the character encoding discerning heuristic:

final String encoding;
if ((tmp.length > 0) && (tmp[0] == 0))

{ encoding = "UTF-16BE"; }

else

{ encoding = "ISO-8859-1"; }

with the following:

String encoding = "ISO-8859-1";
for(int c: tmp) {
c &= 0xff;
if(c==0 || c >= 0x80)

{ encoding = "UTF-16BE"; break; }

}

everything went fine.