Kdtyjb

It seems that newtx is nowadays the best option for Times font, and I wanted to test it with respect to Croatian specifics, but also with respect to some common math symbols, and copying the content from generated PDF.

Consider the following MWE:

documentclass{article}



usepackage[T1]{fontenc}

usepackage[utf8]{inputenc}

usepackage[croatian]{babel}



input{glyphtounicode}

pdfgentounicode=1



usepackage{newtxtext}

usepackage{newtxmath}





begin{document}



a b c d e š đ č ć ž Š Đ Č Ć Ž



bfseries

a b c d e š đ č ć ž Š Đ Č Ć Ž



sffamily

a b c d e š đ č ć ž Š Đ Č Ć Ž



ttfamily

a b c d e š đ č ć ž Š Đ Č Ć Ž



[

  x ne neq y quad

  x le leq leqslant y quad

  x ge geq geqslant y

]



[

  x coloneq coloneqq y quad

  y eqcolon eqqcolon x

]



end{document}

Opening the generated PDF in the latest Adobe Acrobat Reader DC, and copying the content into Unicode-friendly editor, I get this:

a b c d e š đ č ć ž Š Ð Č Ć Ž

a b c d e š đ č ć ž Š Ð Č Ć Ž

a b c d e š đ č ć ž Š Ð Č Ć Ž

a b c d e š đ č ć ž Š Ð Č Ć Ž

x ,, y x ≤≤6 y x ≥≥> y

x BB y y CC x

Found issues:

• I expected both ne and neq to be mapped to U+2260, but they are mapped to U+002C (regular comma).


• Both le (ge) and leq (le) are mapped correctly, but slanted variants are not - they should be mapped to U+2A7D (U+2A7E).


• Both coloneq (coloneqq) and eqcolon (eqqcolon) are not mapped correctly (to U+2254 and U+2255), but to regular B and C.


• Although it looks like đ and Đ are copied correctly, they are not. đ is mapped to U+0111 while Đ is mapped to U+00D0 (not an lowercase-uppercase pair). Considering these two pairs:
  
  
  
  
  • U+00D0 Ð c3 90 LATIN CAPITAL LETTER ETH
  
  
  • U+00F0 ð c3 b0 LATIN SMALL LETTER ETH
  
  
  • U+0110 Đ c4 90 LATIN CAPITAL LETTER D WITH STROKE
  
  
  • U+0111 đ c4 91 LATIN SMALL LETTER D WITH STROKE
  
  
  
  

... I would say that c3*s are used in Icelandic language, while the latter two c4*s correspond to Croatian language. I also checked my source code, and hereby I confirm that my đ and Đ from the keyboard ended as c4*s encoding the source file in UTF-8.

Questions:

• 
  ne and neq: Can someone explain why the mapping is wrong, and can that be improved so they are both correctly mapped?


• slanted variants: Can someone explain why the mapping is wrong, and can that be improved so they are correctly mapped?


• 
  coloneq and others: Can someone explain why the mapping is wrong, and can that be improved so they are correctly mapped?


• 
  đ and Đ: Can someone explain why the mapping is wrong, and can that be improved so they are correctly mapped?

I am very keen to contribute for the improvements for all issues.

Kind regards, Ivan

EDIT: The question about đ and Đ is related to T1 font encoding, not the newtx, as explained in comments below.

glyphtounicode.tex contains a lot declarations to map glyph names to unicode points. But it is not complete. To get e.g. the ne to copy as ≠ add a suitable pdfglyphtounicode:

documentclass{article}



usepackage[T1]{fontenc}

usepackage[utf8]{inputenc}

usepackage[croatian]{babel}



input{glyphtounicode}

pdfgentounicode=1



usepackage{newtxtext}

usepackage{newtxmath}

pdfglyphtounicode{nequal}{2260}



begin{document}

$ne$

end{document}

(I found the nequal by looking in the txsyc.pfb).