Details
-
Type:
Bug
-
Status: Closed
-
Priority:
Major
-
Resolution: Fixed
-
Affects Version/s: 4.3.2
-
Fix Version/s: 4.3.3
-
Component/s: Core/Parsing
-
Labels:None
-
Environment:any
Description
This one came in via support request. The file in question has embedded CID fonts and the toUnicode conversion is handled correctly. However there is a slight hickup related to typographic ligatures and our search system.
When the word "flight" is rendered it is actually converted to "U+FB01ight" and if pasted into a utf-8 encoded looked as it should. However if someone is searching the document for "flight" the match will not be found.
The trick now is how to efficiently updated our text parser to convert Ligatures to their respective "full" unicode values. Values 0x64257 -> 0x64261 are common Latin codes which would need to be converted too two individual character codes.
fi fi U+FB01 fi
fl fl U+FB02 fl
ffi ffi U+FB03 ffi
ffl ffl U+FB04 ffl
ſt ſt U+FB05 ſt
I think this can be implemented fairly easily just need to research the full scope of home many different Ligatures are available.
st st U+FB06 st
When the word "flight" is rendered it is actually converted to "U+FB01ight" and if pasted into a utf-8 encoded looked as it should. However if someone is searching the document for "flight" the match will not be found.
The trick now is how to efficiently updated our text parser to convert Ligatures to their respective "full" unicode values. Values 0x64257 -> 0x64261 are common Latin codes which would need to be converted too two individual character codes.
fi fi U+FB01 fi
fl fl U+FB02 fl
ffi ffi U+FB03 ffi
ffl ffl U+FB04 ffl
ſt ſt U+FB05 ſt
I think this can be implemented fairly easily just need to research the full scope of home many different Ligatures are available.
st st U+FB06 st
Activity
- All
- Comments
- History
- Activity
- Remote Attachments
- Subversion
Marking as fixed.
Show
Patrick Corless
added a comment - Marking as fixed.
I've created a new ligature index in the CMAP class. This map is accessed when the toUnicode() method is called and checks to see if the converted cid matches any of the latin ligatures and will sub in the correct number of respective characters. This address search and text extraction issues and doesn't appear to slow down the content parser in a significant way.