Another week, and I’m still stacking chips. Last week’s chips looked like this:

Intention

At the start of the week, I wrote down four goals for the week:

1. Increase Ezio content buffer — ✅
2. Improve publishing — ✅
3. Fix 404s, implement 301s — both ✅ and ❌
4. Add more old content — ❌

I also added a "stretch goal" of adding more Eurovision data, which once again I ignored.

Update

The very first thing I did last week was the weekchips write up. The second thing was go back to bed, because I lucked out in the side-effects lottery with my AstraZeneca vaccination:

Side effects may include:

• a sore arm ✅
• a headache ✅
• feeling tired ✅
• flu-like symptoms ✅

Heavy sigh.

But then I woke up Tuesday morning feeling normal again, so I only had a 24-hour reaction to it. Tuesday I cracked on improving the publishing on this site, which meant more automatic publishing. I already collect videos that I publish on YouTube overnight, and GIFs that I publish on that site, and I'd like to do this with tweets. But I don't want to archive every single tweet, that seems overkill.

I started with a script that autopublishes based on hashtag, and the first two hashtags I've identified are #itsy and #roxy, our cats.

I added a script that autopublishes based on likes. Not by anyone, only those liked by me using a secondary twitter account (@mnfpublishing) to mark tweets and threads that I would like to archive.

Now, when I import a tweet I try to use the text of the tweet as the title of the post. When it is a thread, such as my trip to the hospital) or just a long tweet, that might not be practical so I arbitrarily decided to cut at ten words. If the first tweet has more, the post ends up titled "Undefined (timestamp)".

I wanted to catch these and other errors before automatically pushing new content to the site. For example, last week I went looking through the logs for 404 errors on the site and found a lot of links to date-based archives such as /2021/4/ instead of /2021/04, which I traced that back to the archive index page. And although I've now fixed that bug in Flourish, having a brake to make sure it doesn't reoccur is a good idea. Plus it gives me a place to inspect for any other problems later.

I did then spend a lot of wasted time debugging that brake, as you can see in Tuesday's activity. To summarise, I was looking for broken links in the generated content using ag -l /...././. And it took me far too long to realise that for whatever reason, the -l flag to ag wasn't reporting filenames. In the end I replaced the silver searcher with ripgrep, and now it works.

Wednesday I finished off chapter 5 of my Ezio content, so (for now) I am running with a two week buffer. I hope to extend that, as occasional pain in my wrists makes it difficult to hold a PS4 controller for long, and I'm trying to keep to a two videos per week routine. I have 14 in the pocket from Horizon Zero Dawn so in the worst case I don't have to make a video for seven weeks, but now I've started releasing Assassin's Creed II I'd like to keep to one of each per week.

I also switched my support scripts to expect .srt format subtitles instead of .sbv, as I discussed in last week's update, tidied up a few other uncommitted changes, and made sure I had backups of important files such as the Davinci Resolve project database.

To further illustrate that whilst I might plan out my week, I don't feel beholden to the plan, on Thuesday and Friday I ignored my goals beyond one quick change to fix the archive page causing 404s.

Instead I spent some time reading and testing how to index the content of my site using xapian so I could add a decent search experience. I have it working on the command-line, and am now building a web interface to the search results.

And then on Sunday, it was the perfect date.