I've got a Literate Agda document, which contains a large number of unicode characters. It builds nicely with XeLaTeX and life is good.
The problem is that arxiv only supports pdflatex, and I'd like to upload the file to arxiv.
Are there any tools/scripts that can process a latex document written for xelatex and add the necessary commands to compile it with pdflatex? Things like declareUnicodeCharacter for each used character.
Basically, I'm just looking for a better way than "start compiling with pdflatex, get an error, add the unicode character, repeat".
Most of the unicode symbols are math symbols, so it shouldn't be too hard to map from the characters to TeX commands automatically.
