Re: [AUCTeX] composed characters in LaTeX source code

[jfbu]

Le 08/11/2018 à 21:10, Stephen Berman a écrit :
> On Thu, 8 Nov 2018 18:58:02 +0100 jfbu <address@hidden> wrote:
>
> > Le 08/11/2018 à 18:45, jfbu a écrit :
> > > Le 08/11/2018 à 18:07, Stephen Berman a écrit :
> > > > I have a file containing such composed characters that I've imported to
> > > > a LaTeX file but in the output of pdflatex, the circumflex is displayed
> > > > over the character following the one it is composed with, e.g., the
> > > > sequence 'b^a' (where '^' means U+0302, the combining circumflex accent)
> > > > is displayed in Emacs with the circumflex over 'b' but in the PDF output
> > > > the circumflex is over 'a'.
> > >
> > > Hi
> > >
> > > I wanted to test your problem but I have another issue, which is that
> > > typing
> > >
> > > 'xb M-x 8 <RET> 302 <RET>ac'
> > >
> > > the 'a' gets superimposed on top of the b in my Emacs buffer:
> > >
> > > (see attached image)
>
> That's strange, and doesn't happen for me (see the screenshot in my
> followup to Joost Kremers).  Does it also look like that in emacs -Q?
> And what version of Emacs?

You are right the problem does not show if using

 /Applications/Emacs.app/Contents/MacOS/Emacs -Q

It is a GNU Emacs 26.1

which is a binary build I got from

https://github.com/railwaycat/homebrew-emacsmacport/releases

(emacs-26.1-mac-7.1)

No problem either in the emacs 22 which I launch in a Terminal window (Mac OS X)

> > > As per your issue, it is going to be very hard in LaTeX to get the accent on
> > > top of previous letter. (I think, but my knowledge of Unicode is
> > > scarce). With LuaLaTeX that could be possible.
>
> Haven't tried LuaLaTeX yet, but XeLaTeX does work, though with
> suboptimal display (see my other followup again).

ok

> > > If really the combining accent is supposedly typed *after*  the letter
> > > (which sounds strange to me, but again, I am no Unicode-guy).
> >
> > Ok, I learned since that's way.
> >
> > As per your original question
> >
> > https://tex.stackexchange.com/questions/79333/is-there-a-way-to-get-pdflatex-to-accept-unicode-combining-accents
> >
> > has a comment by D. Carlisle who said in 2012: I think the answer is "No".
> >
> > A further comment by the same, when asked about "peek at previous character"
> >
> > yes but you can't go back, you can in simple case write a macro that parses
> > the entire text stream re-ordering tokens when it sees a combining character,
> > but it would be very fragile and likely break most other package commands. If
> > your accented letters are single characters in Unicode form NFC then
> > normalising the input before passing to TeX will be a lot more robust.
> >
> > (quote from D. C.)
> >
> > The available answer recommends using Perl to malax the file and normalize the
> > Unicode characters.
>
> Thanks for the URL and quotes; in fact, I had also found that before
> posting, but I don't think it would work for me anyway, because in my
> case it's about composed characters for which there are no corresponding
> single Unicode characters, so nothing to normalize to.

Yes, sorry that I sort of lost sight of that crucial thing,
but then the follow-up idea would still be to pre-process
your file via a Perl script or use Emacs eLisp itself

but now to either output LaTeX mark-up such as you mentioned \^{b}a,

or keep the combining diacritic and \DeclareUnicodeCharacter{0302}{\^},
but move the combining diacritic before the letter.
With the defect it will not look nice in your Emacs buffer.

But you have already considered those options.


I tried XeLaTeX, result was not good with default fonts, but using

\usepackage{fontspec}
\setmainfont{Times New Roman}

it seems to work fine.


When I copy back from PDF I get same combining characters, but
something depends upon from which PDF viewer I copy paste.

- From Skim.app I get like this (where ^ represents U+0302)

xb̂ ac

i.e. there is space after the U+0302

- But from Adobe Reader I get

xb^ac

Of course the first case is bad, because recompiling to PDF
the space does show.

Sorry for somewhat going far away from Emacs matters.

Jean-François