Understanding add@unicode@accent and renewing it for conjoining diacritics

Question

I recently came to know about the command DeclareUnicodeAccent which can declare a diacritic which is printed after the argument of the declared command. e.g.
documentclass{article}
usepackage{fontspec}
setmainfont{CharisSIL}
DeclareUnicodeAccent{foo}{TU}{"031F} % ̟

begin{document}
foo{b}
end{document}

produces b̟ -

Now I want a similar command for conjoining diacritics. It should declare the command in such a way that the diacritic will be placed between the first and the second character of its argument. A dummy (not working) code is as follows.
documentclass{article}
usepackage{fontspec}
setmainfont{CharisSIL}
DeclareUnicodeConjoiningDiacritic{bar}{TU}{"0361} % ͡

begin{document}
bar{ab}
end{document}

Shall produce a͡b -

In source2e.pdf I found the following definition for DeclareUnicodeAccent (which was quite easy to understand), but I didn't understand the definition of add@unicode@accent which must be renewed in order to produce the results correctly.
defadd@unicode@accent#1#2{%
  ifrelaxdetokenize{#2}relax^^a0else#2fi
  char#1relax}
defDeclareUnicodeAccent#1#2#3{%
  DeclareTextCommand{#1}{#2}{add@unicode@accent{#3}}%
}

I want to understand what add@unicode@accent exactly does and how to renew it for the command that I want to develop.

egreg · Accepted Answer

Here's an implementation: documentclass{article} usepackage{fontspec} setmainfont{Doulos SIL} NewDocumentCommand{conjoining}{m >{SplitArgument{1}{}}m}{% makeconjoining{#1}#2% } NewDocumentCommand{makeconjoining}{mmm}{% #2symbol{"#1}#3% } NewDocumentCommand{doubleinvertedbreve}{m}{% conjoining{0361}{#1}% } begin{document} doubleinvertedbreve{ab} doubleinvertedbreve{AB} doubleinvertedbreve{aB} doubleinvertedbreve{Ax} end{document} What does DeclareUnicodeAccent do? defDeclareUnicodeAccent#1#2#3{% DeclareTextCommand{#1}{#2}{add@unicode@accent{#3}}% } It is an interface to DeclareTextCommand, so DeclareUnicodeAccent{foo}{TU}{} is the same as DeclareTextCommand{foo}{TU}{add@unicode@accent{}} Hence, if the current encoding is TU, the command foo{x} will do add@unicode@accent{}{x} and the definition of add@unicode@accent is very simple: defadd@unicode@accent#1#2{% ifrelaxdetokenize{#2}relax^^a0else#2fi char#1relax} which means that if the second argument is empty, the character U+00A0 is inserted, followed by charrelax. It simply isn't what you're looking for.

Marijn · Answer

A small modification of the answer by egreg to lower the accent in case both letters are small (i.e., no ascenders). It works by measuring the height of the argument and calculating the difference between this height and the height of the combination AX.
documentclass{article}
usepackage{fontspec}
usepackage{calc}
setmainfont{DejaVu Serif}
newdimenconjh
newdimenupperh
NewDocumentCommand{conjoining}{m >{SplitArgument{1}{}}m}{%
  makeconjoining{#1}#2%
}
NewDocumentCommand{makeconjoining}{mmm}{%
  settoheight{upperh}{AX}%
  settoheight{conjh}{#2#3}%
  #2raisebox{conjh-upperh}{symbol{"#1}}#3%
}
NewDocumentCommand{doubleinvertedbreve}{m}{%
  conjoining{0361}{#1}%
}

begin{document}

doubleinvertedbreve{ab}
doubleinvertedbreve{AB}
doubleinvertedbreve{aB}
doubleinvertedbreve{Ax}
doubleinvertedbreve{ax}

Huge
doubleinvertedbreve{ab}
doubleinvertedbreve{AB}
doubleinvertedbreve{aB}
doubleinvertedbreve{Ax}
doubleinvertedbreve{ax}

end{document}

wipet · Answer

What you need to do at TeX primitive level looks like this:
defbar#1{barA#1}
defbarA#1#2{setbox0=hbox{#1#2}dimen0=ht0
   ifdimdimen0>1ex advancedimen0 by-1ex else dimen0=0pt fi
   #1raisedimen0hbox{char"0361}#2%
}

bar{aB}

Understanding add@unicode@accent and renewing it for conjoining diacritics

3 Answers

Add your own answers!

Ask a Question