Better than Hugo
In moving the website over, I had to make this obey the new rules over at NearlyFreeSpeech.
One of those involves having to make files end in .shtml if I want to use server-side includes, which are used in the header and footer.
It was a massive pain to get Hugo to just emit files in .shtml, and it's actually hardcoded that using the post-name/index.html file will be prettified to post-name/.
But on a direct string match of index.html, so no go for index.shtml.
So I wrote a Python script that processes my markdown blog posts (relatively few of them at this time) and that seems to get me much better results.
I guess when you hand-roll things…