This question continues one begun in commments to https://tex.stackexchange.com/a/551291/13492.

The issue:

To determine all those glyphs in pdf output, as from the MWE below, that are preventing pdf validation (under standard PDF a 2-u). Namely, because the file fails the rule "The Font dictionary of all fonts shall define the map of all used character codes to Unicode values, either via a ToUnicode entry, or other mechanisms as defined in ISO 19005-2,"

The two commands\pdfcompresslevel=0 and \pdfobjcompresslevel=0 cause the output from pdflatex to be a readable, pure-ASCII file.

The question:

What do I look for in the ascii pdf so as to subsequently include a suitable \pdfglyphtounicode command?

If I process the file with pdflatex first, as it is shown, and then again but with the two lines

    \pdfglyphtounicode{summationdisplay.1}{0060 0060 0060 0060 0060 0060 0060 0060}%
    \pdfglyphtounicode{summationdisplay}{0060 0060 0060 0060 0060 0060 0060 0060}%  

commented out, and if I then examine the resulting ascii pdfs, the only substantial difference I detect is:

  • with those two lines included in the source, the pdf includes once the line

    dup 213 /summationdisplay.1 put

and then the line:

/CharSet (/radicalBigg/radicalbig/radicalbigg/summationdisplay.1/uni222B.dsp)
  • with those two lines commented out, the pdf includes dup 213 /summationdisplay.1 put twice but does not include /CharSet (/radicalBigg/radicalbig/radicalbigg/summationdisplay.1/uni222B.dsp).

Erroneous inference:

From the above, I'm tempted to infer that I will need to include \pdfglyphtounicode commands in the source for glyph names occurring in pdf file lines of the GREP form

dup [0-9]+ /\S+ put

But that surely is not correct! After all, the pdf file includes many such lines, for example:

dup 149 /period put
dup 48 /u1D44E put
dup 49 /u1D44F put

dup 150 /comma put
dup 56 /u1D456 put
dup 58 /u1D458 put

dup 115 /radicalBigg put
dup 112 /radicalbig put
dup 114 /radicalbigg put
dup 213 /summationdisplay.1 put
dup 185 /uni222B.dsp put

dup 61 /equal put
dup 8 /uni03A6 put

dup 33 /arrowright put
dup 49 /infinity put
dup 0 /minus put
dup 184 /plus put
dup 6 /plusminus put
dup 112 /radical put

dup 33 /A put
dup 34 /B put
dup 40 /H put
dup 41 /I put
dup 42 /J put
dup 43 /K put
dup 50 /R put
dup 52 /T put
dup 65 /a put
dup 66 /b put
dup 67 /c put
dup 12 /comma put
dup 68 /d put
dup 69 /e put
dup 1 /exclam put
dup 70 /f put
dup 20 /four put
dup 71 /g put
dup 72 /h put

The MWE:


% To examine pdf as pure ASCII:

      pdfapart=2, pdfaconformance=u,
      pdftitle={A Book}, pdfauthor={Anonymous}, pdfcreator={somebody},
      pdfsubject={A general introducton to things}, pdfkeywords={things, stuff},
    \pdfglyphtounicode{EM}{0058 0058 0058 0058 0058 0058 0058 0058}%
    \pdfglyphtounicode{NUL}{0060 0060 0060 0060 0060 0060 0060 0060}%
    \pdfglyphtounicode{summationdisplay.1}{0060 0060 0060 0060 0060 0060 0060 0060}%
    \pdfglyphtounicode{summationdisplay}{0060 0060 0060 0060 0060 0060 0060 0060}%    
    \pdfglyphtounicode{radicalBigg}{0060 0060 0060 0060 0060 0060 0060 0060}%
    \pdfglyphtounicode{radicalbig}{0060 0060 0060 0060 0060 0060 0060 0060}%
    \pdfglyphtounicode{radicalbigg}{0060 0060 0060 0060 0060 0060 0060 0060}%
    \immediate\pdfobj stream attr{/N 3} file{sRGB.icc}
        /OutputIntents [
             /Type /OutputIntent
             /S /GTS_PDFA1
             /DestOutputProfile \the\pdflastobj\space 0 R
             /OutputConditionIdentifier (sRGB)
             /Info (sRGB)

\newcommand\mytitle{A Book}
\newcommand\myabstract{An introduction to things in general.}



abc abc


The source here is just for experimentation. My real, book-length, document has a pdf ascii output file over a quarter-million lines long! Which is why I need help in what to look for (and what to do about it).



How to find the proper glyph name required by \pdfglyphtounicode

How are the glyph (character) names in PDF-files determined?

How to fix missing or incorrect mappings from glyphtounicode.tex

  • well basically you have to find the names that aren't yet in glyphtounicode.txt or detected automatically). Glyph names like /A or /equal are standard. /uXXXX and /uniXXXX are detected automatically. Jun 29, 2020 at 20:43
  • @UlrikeFischer: OK, that's a big help! (I dare to ask whether trying to make a math-ridden document pdf-compliant is a fool's errand!)
    – murray
    Jun 29, 2020 at 20:46
  • well it would be easier if you would use unicode math fonts. The math times fonts are old and so it is not really surprising that they don't use unicode glyph names everywhere. Jun 29, 2020 at 21:00
  • Does newtxmath qualify as "unicode math fonts"?
    – murray
    Jun 29, 2020 at 22:12
  • no, you would use an unicode font with lualatex or xelatex and unicode-math. Jun 29, 2020 at 22:19


You must log in to answer this question.

Browse other questions tagged .