Undefined control sequence - l.38 \newfontfamily\fixfont about dotpandoc HOT 3 CLOSED

Hi Paul (thanks for mentioning me on the other thread, yes that works!).

I don't use PDFLaTeX but XeLaTeX as my PDF engine and don't really test with PDFLaTeX to be honest. \newfontfamily is a command for XeLaTeX and LuaLaTeX; these engines properly support unicode and more modern opentype fonts.

There has been a regression since last week even for XeLaTeX — I cannot build my workflow.scriv project any more but I could last week (I just updated my sample outputs, see e.g. https://github.com/iandol/scrivomatic/blob/master/sample-output/workflow.pdf updated 5 days ago and working fine). There was a new version of Pandoc (V2.15), and a minor TeX update. I have updated my custom.latex to copy the latest changes for Pandoc, and building now works again...

I can also build your sample. I had to tweak the metadata to use XeLaTeX and add the cleveref package (using header-includes which is pandoc metadata to add text into the latex template header):

pandocomatic_:
  pandoc:
    pdf-engine: 'xelatex'
  metadata:
    header-includes:
      - '`\usepackage{cleveref}`{=latex}'
  use-template:  
    #- paper-with-refs-latex
    - pdf

If I specify XeLaTeX in the metadata, and making minor mods for missing files and replacing your refs with mine it broadly works:

mwe3.pdf
mwe3.md
pandocomatic2.yaml.txt

I normally do not use Pandoc to make the PDF, but output TEX and do it myself (i.e. use paper-with-refs-latex template so Pandoc makes a tex file then I run it through latexmk using scrivomatic automatically with the -B option or manually on the command-line). If I do that, I can fix the minor issue with the first ✉ character:

─❯ pandocomatic --debug -c pandocomatic2.yaml mwe3.md
pandoc	--pdf-engine=xelatex \
	--from=markdown \
	--to=latex \
	--template=/Users/ian/.local/share/pandoc/templates/custom.latex \
	--toc \
	--toc-depth=2 \
	--standalone \
	--filter=/Users/ian/.local/share/pandoc/filters/assimilateMetadata.rb \
	--citeproc \
	--bibliography=/Users/ian/.local/share/pandoc/Core.bib \
	--csl=/Users/ian/.local/share/pandoc/csl/apa6.csl
Pandocomatic needed 1.3 seconds to convert '/Users/ian/Desktop/workflow-mmd/mwe3.md' to 'mwe3.tex'.
─❯ latexmk -logfilewarnings -interaction=nonstopmode -f -pv -time -xelatex mwe3.tex

mwe3.pdf

Now if I specify pdflatex as the engine, I get this:

mwe3.pdf

You see the first page has a minor error but that can be removed; indeed I have now put these lines in an IF statement so they shouldn't be called anymore if using PDFLaTeX... You can upgrade the custom.latex template...

from dotpandoc.

Some notes:

1. You should use nonumberfigures: false in your metadata — I set it to true as I use Scrivener to generate figure numbers and my cross-references.
2. Not sure why you use PDFLaTeX, but note - postprocessors/fixLaTeX.rb will inject \fixfont for certain characters as it assumes you are using XeLaTeX or LuaLaTeX; so you will probably want to disable this postprocessor. See https://github.com/iandol/dotpandoc/blob/master/postprocessors/fixLaTeX.rb#L12 for the details...
3. You can add more packages using header-includes — or you can edit custom.latex and make your own template to include these. I think header-includes works well for simple additions.

from dotpandoc.

Indeed, using --pdf-engine=xelatex did the trick. I'm not sure why I used the pdflatex-engine; I think it is reminiscent of my code from 10yrs back or so.

I'll close the issue!
Thanks for your support.

from dotpandoc.