I do wish PG would spend more time making his language better and it does feel a bit like he's abandoned it. If Clojure compiled to C instead of sat on top of Java, I would probably jump ship. Having a community is a valuable component of a language and PG's in a great position to really market Arc (we already have proof that he can build a community due to HN), but he just doesn't. My conclusion is that running YC has caused him to strongly deprioritize Arc.
While I still hack in Arc constantly because I'm always tweaking HN, I haven't made dramatic changes to the language itself lately. I only seem to be able to work on 2 things at once, not 3. Since YC is a given, that means I have to choose between hacking and writing. Lately I've mostly chosen writing.
I should release a new version though. News.arc is greatly improved since the last release.
That would be great. Seeing others share code - even tools that I don't currently use - makes me want to push code as well. When I push code and no one else does, I feel a bit like my work is unappreciated and want to stop sharing. I suspect there is type of momentum at work in the formation of vibrant programming communities. Seeing Arc 3.2 would be motivational.
iirc all reported bugs fixes were incorporated by the arc3.1 release. (Releases haven't always incorporated all fixes reported up to that point, the atomic-invoke fix was particularly alarming and took several releases to make it in).
The only bug I'm currently aware of off the top of my head in arc.arc is readline, which was reported after the arc3.1 release.
There are a couple of known issues with the Arc runtime (i.e. the queue bug you found which seems likely caused by unsafe mutation of immutable cells, nested quasiquotation) which have prospective fixes, but neither tested throughly enough that I'd personally say, "oh, why yes, you should go ahead and switch HN over today".