Jan-Lukas Else

Jan-Lukas Else

Tech, life and everything else

2020


Find broken links on your website with muffet

Published on in 👨‍💻 Dev

Today I wanted to check if my blog contains some broken links and it actually did. I used the Go-based tool muffet which crawls a website and checks for HTTP errors when following links. To speed up the process, I started a local Hugo server and used the following command for muffet (after installing it):

View

Published on in 💬 Micro

One way to start the day well: a croissant for breakfast. 🥐

🖼️ View

Using the Google Text-to-Speech API to create audio versions for blog posts

Published on in 👨‍💻 Dev

Ashwíṅ Víshnú asked how I created the audio version for my latest post. Here’s a way how you can use the Google Text-to-Speech API to create MP3s from text. A simple way to try this API without a Google Cloud account is to follow these steps:

View

Attention

Published on in ✍️ Posts
Updated on

When was the last time you watched a 90 minute movie without drifting off with the thoughts, pulling out the smartphone or doing something else in between than watching the movie? I often cannot even manage to watch a 30 minute video without being distracted.

View

Pure CSS paintings

Published on in 🔗 Links

Ben Evans created amazing pieces of art. He “draw” a sunset, a still life and a portrait. But instead of images, he created CSS style sheets. All those images are made with just CSS. It is very impressive!

View

Writing a SQL database from scratch in Go

Published on in 🔗 Links

Phil Eaton wrote a really cool article. He documented how he used Go to write a rudimentary SQL database including a CLI. It is very interesting to read how commands are parsed and analyzed. The whole thing probably has no practical use (there are already countless mature database systems), but it’s still exciting. I remembered some of the math lectures in the first semester about formal languages.

View

Published on in 💬 Micro

No, my implementation (it’s still a WIP) doesn’t support sending “Update"s and “Delete"s yet. It’s one thing that bothers me about ActivityPub. Webmention is so much simpler in that regard. There also seems to be no way of forcing Mastodon to refetch an activity object other than sending an “Update” or “Delete” activity.

View

Published on in 💬 Micro

Just a small note: I deleted my Microblog.pub instance. I liked this piece of software and it’s a nice single-user alternative to Mastodon, but I don’t need a Fediverse instance anymore. My new fediverse strategy is to use my normal blog to post stuff (you can follow me via @en@jlelse.blog and @de@jlelse.blog) and Miniflux to follow people (Mastodon offers RSS user feeds).

If you reply to posts or mention my user handle in a public post on Mastodon, I will receive a notification and your reply will appear in the “Interaction” section under the post. If you interact with a post on your website and don’t automatically send Webmentions, you can also submit the URL in the form below each post.

By removing Microblog.pub from my server it also reduces the need for resources a lot. Python as well as MongoDB don’t seem to be that resource friendly. I currently use the cheapest VPS option from Hetzner and for the few remaining things hosted there (my website, my blog and some Go tools etc.), it’s more than enough.

View

Careless neighbors

Published on in 💭 Thoughts

I live in an apartment building with neighbors in all directions. But some neighbors don’t seem to care at all whether they disturb their neighbors with very loud music. At the moment I don’t have to study, but with this noise I couldn’t study at all. It’s not an exaggeration, but I hear the neighbors’ music as loud as if I were listening to music myself either through speakers or headphones. By now I have come to terms with the almost daily nightly screaming of their child, which wakes me up as a result. Maybe the child is in pain or something? But I find this loud music simply outrageous.

View

Published on in 💬 Micro

I use BunnyCDN as storage for my Micropub media endpoint, where I upload all the photos and other media files I publish on my blog. When I upload a new photo the media endpoint first uploads the original file with the name of it’s SHA-256 hash value (so when I upload the same file multiple times, it doesn’t create multiple files). When I upload an image (JPEG or PNG) it also gets optimized and resized to a maximum width of 2000 pixels using the Tinify API (and then uploaded to BunnyCDN too), doing that significantly reduces the file size of the images displayed on my blog.

It would be probably awesome to also serve different formats based on the Accept HTTP header (WebP for browsers that support it) or different sizes, but I decided I want to hold my perfectionism (and the complexity) in check and settle for using just one optimized file version in the original file format. If at some point all browsers support WebP, I might consider using only WebP files, but until then it still takes time.

At the moment I don’t backup the files from BunnyCDN (I trust that they - or I - don’t accidentally delete them), but I could probably setup a cron job to regularly copy new files via FTP from their storage to another storage.

View

Jan-Lukas Else