Unicode characters are not handled properly in LaTeX exports

[johnblommers]

I've placed the characters π (π) and → (→) in a novelWriter document. When I export to LaTeX and attempt PDF conversion using TeXshop two errors present themselves:

./Dick_and_Jane.tex:66: Package inputenc Error: Unicode character → (U+2192)
(inputenc)                not set up for use with LaTeX.

See the inputenc package documentation for explanation.
Type  H <return>  for immediate help.
 ...                                              

l.66 Jack and Jill →
                       the hill to fetch a pail of water.
? 

./Dick_and_Jane.tex:68: Package inputenc Error: Unicode character π (U+03C0)
(inputenc)                not set up for use with LaTeX.

See the inputenc package documentation for explanation.
Type  H <return>  for immediate help.
 ...                                   

l.68 Puff likes to eat π
                          just like humans.

The two unicode characters don't appear in the PDF output after we press Return twice to bypass the error messages.

I had thought the matter resolved with the addition of the latexcodec but I think I've jumped the gun and tested this prematurely.

[vkbo]

The latexcodec documentation states that it isn't complete. I suppose all we can do is identify the problems and add them to the replacement dictionary manually, or report them to the package developers.

I will add these two, and check a few others of similar type.

[johnblommers]

I've found a workaround. If the intention is to export to LaTeX then the writer can simply insert the LaTeX code for the symbols they want. So:

• PI can be presented as $\pi$
• RIGHTARROW can be represented as $\rightarrow$

novelWriter already passes the LaTeX codes directly into the LaTeX output file where it will present correctly when converted to PDF. I'm a Gummi fan and this worked for me.

[vkbo]

Yes, the latexcodec package leaves already correct LaTeX formatted text alone.

Only paragraphs and heading text is encoded with this package though, but this is certainly a workable solution if LaTeX is the only export used. I will add it to the documentation.

[johnblommers]

Comment.

Writers that include LaTeX in their novels are probably a rare breed and will know how to handle problems they encounter using it in novelWriter.

For example. I created a LaTeX math formula that contained subscripts like this:

$$f(x)=a_{0}+a_{1}x$$

but on exporting this with novelWriter the subscript character _ found itself escaped as \_ in the .tex export file. The workaround is to instead use the \sb LaTeX syntax:

$$f(x)=a\sb{0}+a\sb{1}x$$

[vkbo]

Yeah, that's far beyond what is intended usage of novelWriter. It's for writing plain text. The LaTeX export is only intended for generating PDFs of plain text quickly.

[vkbo]

LaTeX export will not be included in the new Project Build Tool, and the current Export tool will be removed. This issue will therefore not be addressed. Closing it.