because this web site is for humans. Not that it will matter since the
robots just ignore this altogether now, but I want to make it clear
they're not welcome. Might implement a way to poison the content for
them down the line.
I'm moving to self-hosting this site and building it from my self-hosted
code repo and CI, which won't have access to my FreeTube subscriptions
when it builds. To get around this, I've moved to a script that will
snapshot my local FreeTube subscriptions to a JSON file. I'll commit this
file, and it can be used in the build on CI.
I thought the channel names were already sorted in the source data file,
but it turns out they aren't. I now sort them after sorting the category
names.
Animated gifs can't go through the asset pipeline since they come out as
a single frame of the original animation. Previously, I had only a
couple of them and didn't have my assets structured very well. I had the
assets that needed to be passthrough copied alongside those that could
go through the pipeline. It wasn't bad since there weren't many
animations. I mapped just those to passthrough copy to go to a single
directory. This will become more fraught as I add more since there's
more opportunity for collisions. When mapping to an output directory, I
can't maintain the inner structure of that directory. The files are
flattened. I moved everything around, both to separate images that can
go through the pipeline from those which can't and to allow for a simple
passthrough copy that will maintain the directory structure so that each
blog post can have its own directory of passthrough copied animated gifs
in the build.
I want the home page to use just `title` as the page title, but I want
other pages to add the site name to the page title. This filter does
that and spits out the final title.
