Responding to second round of feedback on my Ionmonkey project: done! Now with five patches where one would do ;-) (Somewhat reassuringly, when I implemented the test cases that Jason suggested, not all of them worked the first time -- in particular, I had to change toString() for functions to pretty-print generators as function*.)
(This time I finished with that at 4:15PM rather than 10:15, which is good.)
In other news, I tried to make a snapshot, and have not succeeded so far since the past few days have been very sad days for our incoming branch and there have been many broken builds to get through before getting to the snapshot build in the queue. Plus, the Linux bot apparently got wedged trying to run valgrind on something, or at least, it ran for more than ten minutes before I killed the process. So now I'm just waiting for the FreeBSD bot to finish running, and I can push a new snapshot that fixes a trait-related bug I found, which means I can push a change to the core::num library that means one less FIXME in the compiler... oh, the slow progress of compiler work.
Back to removing match check... which is currently blocked on the yak-shave that I mentioned before, which is causing mysterious assertion failures even after rewriting the change from scratch. It seems to be happening in pretty-printer code that the macro-expander is calling, which means it's something I don't want to deal with. Maybe I'll just scrap that change for now and take the easy way out (inserting more failure cases) so I can make progress).
I took a break after writing the paragraph about the snapshot, and in the meantime, the snapshot finished building and I got it, along with the traits change, checked in. And science marches on.