TeXForm of 2 is not supported.

[vsht]

I started to transition the full FeynCalc documentation to Markdown using M2MD not so long ago,
but there are still some edge cases where I cannot obtain a proper LaTeX output and get only an image instead.
For example

M2MD[Cell[BoxData[
    FormBox[
      FractionBox["1", 
        TemplateBox[{"\"(\"", 
            SuperscriptBox[
              FormBox[
                FormBox["p", TraditionalForm], TraditionalForm], 2], 
       "\"+\"", 
            RowBox[{"2", " ", 
                RowBox[{"(", 
                    RowBox[{
                        FormBox[
                          FormBox["p", TraditionalForm], 
              TraditionalForm], 
                        
             FormBox["\"\[CenterDot]\"", TraditionalForm], 
                        FormBox[
                          FormBox["q", TraditionalForm], 
              TraditionalForm]}], ")"}]}], 
            RowBox[{"-", 
                SuperscriptBox["m", "2"]}], "\"+\"", 
            RowBox[{"\[ImaginaryI]", " ", "\"\[Eta]\""}], "\")\""},
          "RowDefault"]], TraditionalForm]], "Output",
   CellLabel -> "Out[13]="]]

During evaluation of In[15]:= TeXForm::unspt: TeXForm of 2 is not supported.

Out[15]= $$![1n6s7udnvjhyu](img/1n6s7udnvjhyu.png)$$

This happens for SFAD[{{p, 2 p . q}, m^2}] in FeynCalc. Is there perhaps a simple workaround for this problem?

[vsht]

It seems that setting

System`Convert`TeXFormDump`maketex[i_Integer] := ToString[i]

fixes the issue in this example. I'll check if it helps in all cases and report back.

[vsht]

Works for all my examples.

[szhorvat]

Have you complained to WRI about TeXForm? They are more likely to fix things when more people complain.

My biggest complaints:

• It hasn't been updated in ages, and does not support many "newer" construct ("newer" meaning introduced in version 10.0 in 2014 seven years ago ...) Just try TeXForm[<|1 -> 2|>], and consider that associations are a fundamental data structure now.
• It keeps outputting \unicode{1234} for unsupported characters. Unfortunately, this is next to impossible to make interpretable either by LaTeX or MathJax (even ignoring the fact that most of these point to private character codes)
• It is not customizable by users, so neither can we update it to work with our own stuff, nor can be fix any of the glaring omissions!

Sometimes I suspect like I'm the only one who keeps complaining, so they assume that it's "just the one grumpy user"

[szhorvat]

@vsht I also try to generate the IGraph/M documentation using M2MD and Pandoc, see http://szhorvat.net/mathematica/IGDocumentation/ I am very interested in sharing ideas. I have not yet managed to make multi-page documentation work.

[vsht]

Have you complained to WRI about TeXForm? They are more likely to fix things when more people complain.

Yes, I did. It would be great if they could at least fix this weird TeXForm of 2 is not supported bug, since the fix is trivial.

@vsht I also try to generate the IGraph/M documentation using M2MD and Pandoc, see http://szhorvat.net/mathematica/IGDocumentation/ I am very interested in sharing ideas. I have not yet managed to make multi-page documentation work.

I've seen your IGraph/M documentation and it indeed looks very nice. My plan is actually to replace

https://feyncalc.github.io/FeynCalcBook/guide/FeynCalc.html

using M2MD, but it still might need some work. Furthermore, I'd like to have FeynCalc documentation available
as a PDF file, where the Markdown output of M2MD will be converted to LaTeX via pandoc.

The good thing about M2MD is that one can put TeX (with $) directly into .m files like in

https://github.com/FeynCalc/feyncalc/blob/master/FeynCalc/DocumentationFiles/Mathematica/Dirac/DiracOrder.m

and it gets properly processed. At the moment I'm still in the process of polishing the .m files that I automatically
generated from our documentation notebooks.