It was a long day. I was lurking HackerNews and /r/programming. Things were going slow. I had been taking quite a lot of notes inside a plain ol’ directory. I know, 2016 right?! But every single note taking application had failed me.
Although I had no use of viewing my markdown notes in the browser, I decided to do it anyways. For the sake of completeness.
I immediately thought of dystic, my personal static site generator, but dissed it as too heavy. I decided then, that I wanted something I could hack together in the next 10 minutes.
I immediately thought of pandoc.
pandoc --toc --from markdown --to html my-note.md -o my-note.html
Done. Easy, wasn’t it?
Now let’s do this for the entire directory, recursively. We could write a simple shell script, but that would run everytime on everyfile. We don’t want that. We want to regenerate the html only if the markdown counterpart has been updated since the previous build.
better faster way than to use GNU Make for
cd /notes nvim Makefile
# Find all markdown files MARKDOWN=$(shell find . -iname "*.md") # Form all 'html' counterparts HTML=$(MARKDOWN:.md=.html) .PHONY = all tar clean all: $(HTML) %.html: %.md pandoc --from markdown --to html $< -o $@ tar: $(MARKDOWN) tar --exclude=notes.tar.gz --exclude=.git/ -czvf notes.tar.gz ./ clean: rm $(HTML)
We’re almost done. Now just run
make all in
/notes and all your
markdown files will be built into html files.
make clean to remove all the html files and run
make tar to
backup all your notes.
Simply open your browser and type “/notes” (or your complete
path) into the address bar and voila! It’s not pretty but it
works. (PS: use
file:///path/to/folder if not using chrome)
To make your files more pretty use my minimal css theme sakura and then change the pandoc command as follows:
cd /notes wget "https://raw.githubusercontent.com/oxalorg/sakura/master/sakura.css" pandoc --css /notes/sakura.css --from markdown --to html $< -o $@
Now you can remote sync the entire website easily using rsync:
rsync --exclude '*.md' source/ destination/
That’s the basics. This can easily be used to create your own blog, websites, small projects, pretty much anything. The best part being that it needs no ‘rules’ from your end, you’re free to structure your content anyway you like; something which is missing from almost every static site generator out there. This is one of my quibbles which I’m trying to fix with dystic.
I added a couple more feature including making automatic
indexes, sorting using title/date, metadata parsing etc. But I
quickly realised that it’s a lot of pain to be doing it using
only GNU Make. So I’ve decided to start working on
again. Maybe even re-write it in
Let me know your comments below.