« Day 4 | Day 6 »

Today I am spending time on investigating of other items from my TODO with code experiments and reading through their reasoning. At least, I’ll just understand them a bit better (to use in the future), and at most, I might put some examples/explanations into the changelog draft.

So, I am starting to read through WeakKeyMap justification ticket and play with it and WeekMap to understand them better.

(Skip 40 minutes of reading the discussion, docs and playing with code… My ADHD is well-compensated, so I tend to just call it “my method of the problem studying”, even if it doesn’t move me much closer to writing a changelog entry.)

A few random notices:

  • It turns out that despite similar names, WeakKeyMap’s and WeakMap’s APIs are quite different (the new class, say, doesn’t have any iteration methods, as they are hard to implement correctly with weak references)—which breaks my assumption from Day 3 (as a reason to improvie WeekMap’s docs during changelog work): “most of the methods should’ve been similar!”;
  • There are internal differences, too: WeakKeyMap compares its keys by equality (akin to a regular Hash), while WeakMap compares them by identity;
  • WeakMap’s docs (here’s 3.2 version) have a lot of small problems: like, both each and each_key have a description “Iterates over keys and objects in a weakly referenced object”, and each_value is explained as (emphasis mine) “Iterates over values and objects in a weakly referenced object”; so, it seems that improving them is a good thing to do this year, anyway?
  • WeakKeyMap#getkey has a confusing at first protocol “just return the key if it exists,” but deduplication use case in the original discussion seems to clarify that.

Next, I investigate a Thread::Queue#freeze prohibition discussion, which is— quite lively, let’s say; but the investigation is also fruitful, as by the end of the discussion there is a clear explanation why freezing the Queue should be prohibited (and not just fixed to forbid #push and #pop, which was the original ticket’s scope).

And that would be it for today, no new commits, yet a few things became clearer!

« Day 4 | Day 6 »