git-diff-pdf.md

Comparing PDF files in git diff commands

The typical approach to comparing PDF files in git diff outputs amounts to converting PDF files into texts through pdftotext and show the diff of conversion.

1. Install the utility pdftotext from the project poppler
2. Enable the handler pdffiles in diff for PDF files and instruct the handler to call the pdf-astextplain script:

   echo "*.pdf diff=pdffiles" >> ~/.config/git/attributes
   echo "[diff \"pdffiles\"]\n\ttextconv = pdf-astextplain\n\tbinary = true" >> ~/.gitconfig
   

   See: https://git-scm.com/docs/gitattributes#_marking_files_as_binary
   and https://git-scm.com/docs/git-diff#Documentation/git-diff.txt---textconv
3. Add the wrapper of pdftotext to direct its output to stdout. Create a script pdf-astextplain in $PATH:

   #!/bin/sh
   pdftotext -layout -enc UTF-8 "$1" -
   

   To compare metadata in addition to the content of the PDF, add pdfinfo "$1" in pdf-astextplain.
4. In some GIT implementations, there is a astextplain script that converts PDF and other files to text for diff as well.
   See: https://github.com/git-for-windows/build-extra/blob/c223c7757745c1df552c0dd4628c368aaea11f32/git-extra/astextplain
5. In similar spirit, use zipinfo -l to show contents of ZIP archive and use the script racket-wxme-astext to show Racket WXME files.

pdf-astextplain

#!/bin/sh
pdfinfo "$1"
pdftotext -layout -enc UTF-8 "$1" -

racket-wxme-astext

#!/bin/bash

if ! file --mime "$1" | grep octet-stream > /dev/null; then
  cat "$1"
else
  cat "$1" | racket -n -l racket/base -l wxme -l racket/port \
    -e '(copy-port (wxme-port->text-port (current-input-port)) (current-output-port))'
fi

zip-info-astext

#!/bin/bash
zipinfo -l "$1" || unzip -l "$1"