When Cyberpunk 2077 was announced, I was really excited. The teaser for it looked exactly like the Cyberpunk 2020 RPG that I’d like so much when I was a kid. When it came out, I tried it and it was a mess. Later, I heard they worked out lots of the problems, and I went back and I’ve been playing it. It still looks just about perfect, and gets lots of things right. It feels like a really good adaptation of Cyberpunk 2020. There are a lot of bugs and interface issues, still, but I don’t want to write about those, because it’s boring.

A few months ago, I wrote about a new 1Password library for Perl, which I was using to stop putting sensitive information into my environment. I was pretty happy with this! It meant I could put a pointer to my credentials in my configuration, instead of the credentials themselves.

I used to program on paper, then type it in later. Not all the time, but sometimes. Sometimes I’d write pseudocode. Sometimes I wrote just the code I would type in later. Sometimes just flow charts and subroutine signatures. These days, I only really do the last version, because I always have a computer nearby now. I’m not stuck in a boring lecture, for example, with only a legal pad.

This post will be short but baffling.

The other day, a concatenation of circumstances led me to thinking about the lousy state of sending formatted text to Slack. We have a bot called Synergy at work, and the bot posts lots of content. Mostly it’s plain text, but sometimes we have it send text with bold or links. This is for a couple reasons. Our bot supports channels other than Slack (like SMS and Discord and the console), so we can’t express everything in Slack-oriented terms. But even doing so would be hard, because of the APIs involved.

A couple weeks ago, I had to test a printer, and the first text file I found to test-print was this, quazza-purchases.txt:

Quite a while ago, I wrote some Spotify code that would find places when my Discover Weekly playlist would intersect with those of my friends and coworkers. This was fun. Every once in a while, I talked about other things that might be fun, along those lines. I tend not to do too much with those ideas, because the Spotify API is often missing exactly the method I want, and I can work around it, but it becomes too much of a pain. Also, I’m a little lazy when the project will need other people to be interested, unless I’m sure they will be. Who wants to launch a flop to their friends?

I’ll be quick, this is just an update to my last post!

In my last post, I wrote about how I made dzil workflower to install GitHub Actions into my Dist::Zilla-based dists’ repositories for automated testing. I also said I’d been reading O’Reilly’s Learning GitHub Actions. This week, I applied some more of what I learned from the book, and it was good.

Starting ages ago, once in a while somebody would show up and offer me a commit that would add some automated testing system to my open source repositories. I liked the idea, but it always felt like a free puppy. I didn’t know how it worked, I didn’t know how the YAML file (always YAML!) was put together, and I didn’t know what I was supposed to do when something went wrong.