Today, I am just continuing the regular paint-the-fence work of enhancing the formatting, structure, and examples throughout the file.

It turned out not that much left to handle, in principle, though there are still questions, uncertainties, and challenges, like:

  • It is hard to add a persuasive example for String#bytesplice with the new protocol (copy part of one string into the other), because this specific change is most useful for implementing low-level stuff like code editors and network protocols; so I just add a demonstrative ones, with simple strings (with Ukrainian text, just to demonstrate Unicode, you know).
  • I add base example/demo for Fiber#kill, but plan to do more, demonstrating its edge cases and limitations; I always feel kinda weird documenting fibers: this is a piece of well-designed machinery, easy to understand and explain and able to achieve complex things— which I happen to never use in my code (which is mostly application-level/business-level, usually, while fibers are a powerful underlying abstraction for enumerators and async code).
  • I try to show some code examples for Process::Status deprecated methods… and decide against it, eventually: it seems that concerned parties (rare, by my estimation) wold anyway understand better what the deprecation means and how to check it;
  • I add a bit more details about Prism parser and notice that its docs are rendered quite badly by RDoc on the Ruby Core docs site. (On its own site at least method docs are informative, while module docs are stil taken from some auto-generated files.) It is frequently problematic with big and complicated libraries, especially when they are integrated into the main Ruby’s RDoc. I put into my “long TODO of Ruby docs” to look into it some time in the upcoming months.
  • I start to link the standard library names to the corresponding repos… And think that some time in the future I’d better automate that. Probably there should be a machine-readable data about current versions, repo links and default/bundled/non-gem status of each. And, once I think about it, I understand that Jan Lelis’ awesome stdgems.org probably should have a machine-readable version… And indeed, it does! So I probably should somehow incorporate that into Ruby Changes build process (though not this week, probably!)
  • Which also makes me think about the fact it would be cool to have a machine-readable changelog for each and every Ruby library (and, for that matter, for Ruby itself).
  • Which reminds me of a long-lasting discussion/plan about making Ruby Changes itself into some formal DB/YAML, which would probably open some interesting possibilities. Though, I am not sure how/when this avenue of thought will be followed.

Anyway, I finally fill “Highlights” section (again, confirming that this version’s language/API changes are smaller than some previous versions), and with that— it is mostly ready? Commit: 7f65d0f

There are still a few TODOs here and there, and I should think about some better description at least for Process.warmup, but overall—

Let’s say even if for some unexpected reasons I wouldn’t be able to work on it anymore, it can be published already, with a couple of small fixes.

So if I will have time to continue this advent, I plan to spend a next few days on general housekeeping: both for the changelog’s site improvements and on Ruby’s docs/NEWS.