I did myself a code yesterday!
I mean (puts on professional mask), I finally found and fixed a bug in Podcast chapters which would cause some problems with comment frames. It is the kind of problem which, depending on your workflow, most people never see. But those who do see it, they see it precisely all the time. I have not got a build submitted yet, because it feels good to test this out on slightly more data than just the things my own app has generated.
Once the fix is in place and feels solid, I think it points toward some very nice cleanup opportunities for the code overall as well. In brief, reading and writing text stored in the file metadata has some intricacies to it, and Podcast chapters has mainly hacked around it on a case by case basis. In most places, knowledge of the data you are writing lets you get away with this, but the comment bug happened to point me toward a much more correct way of doing things. I still use a horrible hack to actually do the right thing at the moment, but that is nice and encapsulated so that I can clean it up once I know everything does what it should.
The solution might even be pointing toward a cleaner solution for my whole system of reading and writing the actual bytes of metadata. I have always felt that it is clunkier than it should need to be, so that is an exciting thought.
On links on Linkedin
I, of course, had to post a link to Linkedin about yesterday's post on how nobody follows links on Linkedin. The theory still holds, the page got virtually no views at all. So if I wanted to do any more playing with poking the networking bear, I guess it would be posting a whole post right there, with or without a link to here (the real home of my texts).