For reasons, I grabbed the entire MELPA archive of published Emacs packages to grep against. Some questions are answerable only by inspecting existing code.

Fetching MELPA was staightforward — it’s documented in their README.

rsync --info=progress2 -avz --delete rsync://melpa.org/packages/ snapshots/
# then unpack it
mkdir x
for tarball in snapshots/*.tar; do echo -n .; tar -C x -xf "$tarball"; done; echo

A quick wordcount later:

find x -type f -name '*.el' -exec wc {} + | tail -1

yields:

909855  3647709 39331411 total

Written for humans:

• About 910k lines
• About 39 Megabytes of Elisp source

Zooming out, there’s far more data in MELPA than code:

find x -type f -exec du --human-readable {} + | sort --human-numeric-sort -r > du.txt && head du.txt

Shows only two of the top files are actually Elisp source:

There is music in MELPA.

Who would have known.