Why doesn't pandoc convert a plaintext file to PDF properly?

Question

Commands tried:

pandoc -V 'fontfamily:Courier' --variable mainfont="Courier" --pdf-engine=pdflatex 1.txt -o 1.pdf
pandoc -V 'fontfamily:Courier' --variable mainfont="Courier" --pdf-engine=lualatex 1.txt -o 2.pdf
pandoc -V 'fontfamily:Courier' --variable mainfont="Courier" --pdf-engine=xelatex 1.txt -o 3.pdf
pandoc -V 'fontfamily:Courier' --variable mainfont="Courier" --pdf-engine=latexmk 1.txt -o 4.pdf
pandoc -V 'fontfamily:Courier' --variable mainfont="Courier" --pdf-engine=tectonic 1.txt -o 5.pdf
pandoc -V 'fontfamily:Courier' --variable mainfont="Courier" --pdf-engine=wkhtmltopdf 1.txt -o 6.pdf
pandoc -V 'fontfamily:Courier' --variable mainfont="Courier" --pdf-engine=weasyprint 1.txt -o 7.pdf
pandoc -V 'fontfamily:Courier' --variable mainfont="Courier" --pdf-engine=prince 1.txt -o 8.pdf
pandoc -V 'fontfamily:Courier' --variable mainfont="Courier" --pdf-engine=context 1.txt -o 9.pdf
pandoc -V 'fontfamily:Courier' --variable mainfont="Courier" --pdf-engine=pdfroff 1.txt -o 10.pdf

Contents of 1.txt:

--------------------------------------------------------------------------------
Left                                                                       Right
--------------------------------------------------------------------------------
Lorem ipsum whatever. Lorem ipsum whatever. Lorem ipsum whatever. Lorem ipsum  1
whatever. Lorem ipsum whatever. Lorem ipsum whatever. Lorem ipsum whatever.    2
Lorem ipsum whatever. Lorem ipsum whatever. Lorem ipsum whatever. Lorem ipsum  3
whatever. Lorem ipsum whatever. Lorem ipsum whatever. Lorem ipsum whatever.    4
Lorem ipsum whatever. Lorem ipsum whatever. Lorem ipsum whatever.              5
--------------------------------------------------------------------------------

Results:

Out of all those allegedly supported "engines", only the first and third produce any PDF at all (the others just dump a bunch of nonsensical errors). And those two that do produce PDFs, produce horribly butchered ones:

1. "pdflatex" (the first command) entirely ignores the specified font, so it's completely useless.
2. "xelatex" (the third command) seems to be mostly using the right font, but seemingly deletes all the spaces between "Left" and "Right", morphs the "-"s into straight lines (that's not how that font looks...) and messes up the lines completely so that the numbers on the last columns are not aligned to the right, and has crammed the entire contents into the middle of the page instead of, as expected, near the top-left corner:

screenshot of the xelatex-produced PDF

I have spent enormous amounts of times hunting for options and trying a million variations of the above commands, but it seems like this tool is fundamentally broken. I have no idea how others (apparently) use these tools, but they just don't work. It's impossible to convert a text file to PDF...

Answer 1

Pandoc is not broken; it is doing just what its documentation says it will do. Pandoc treats your input file as Markdown with pandoc extensions (since you didn't specify a format). What you have here is a one-column simple table (since there is no break in the line of ----s to indicate a column break).

If what you want is a rendering of this context as verbatim text in a PDF, you could use e.g. enscript 1.txt --output=- | ps2pdf - > 1.pdf. If you want to do it using pandoc, then the easiest way is to put the content inside backtick fences so that it is treated as a markdown verbatim block. One way to do this would be to modify your file, but you could also do it by creating a file ticks.txt containing just

```

and then run

pandoc ticks.txt 1.txt ticks.txt -o 1.pdf