Yes, that’s correct, and that’s what I see in pdf. When I use nodes.toutf I suspect that now output in log file is no longer decomposed in T + h chars but it’s the Unicode char corresponding to Th ligature which editor font does not have glyph for.


Apparently in previous versions the fonts.hashes.characters lua structure used to store unicode of components of ligature (so unicode element was a lua table/array) and now its 1 single value.


local fonthashes = fonts.hashes

local chardata = fonthashes.characters


local info = chardata[font][c]


Before, info could have been a lua table (having 2 values, unicode for T and unicode for h) or a number (unicode for some single non ligature). Now it seems to be always 1 single value (unicode for Th).

Is there anyway else I could get the decomposition of the ligature ?





De : Rik Kabel
Envoyé le :mercredi 27 février 2019 18:37
À :
Objet :Re: [NTG-context] Output problem with nodes.toutf utility function(latest beta 02.26.2019)


On 2/27/2019 11:53, Joseph Canedo wrote:

Dear list,


I use lua function nodes.toutf to print text being typeset in log output for debugging purposes etc… and I noticed output is now different for some characters/glyphs.

I use EB Garamond font, which might explain.


Before, for example output was :


user            > margin > margin box text Thareh.


Now it is :


user            > margin > margin box text 󰀙areh.


The « Th » was replaced by some 󰀙 placeholder perhaps for unknown character?


Please note that typeset text is not impacted, it seems correct.




The dlig feature for EBGaramond causes "Th" to be replaced by a ligature. Look carefully at your typeset text and you will likely see that the "T" and "h" are joined.