Rondo Dondo

When the last of the black holes evaporate I will be there to remember where the name "nopert" came from

Behold, my beautiful proc macros that let me avoid importing the maybe_async crate

I believe the more relevant issue is that the Arc that sees the ref count drop to zero is the one that drops it. Thus, if you cannot guarantee that the drop occurs on the same thread that the inner type is locked to

Damn that game ripped so hard

Not only that, but into_inner returns the inner type if reference count is 1

The year 3000

😂

Zohran is kinda an economic populist imo

Despite my best efforts, I've become fewer-dependency pilled

I think it’s about time scheme-rs no longer requires async

My next potentially biggest optimization relies on the following question: which is faster, a non blocking recv from a channel, or mallocing a value?

How did I know instantly that this picture was from the wikipedia article?

I just need to make my call/cc implementation about 8x faster and I'll be happy.

Aw yeah, that's the stuff

Just now realizing another term for zombie with no conscious would be a philosophical zombie

I have a feeling this next scheme-rs optimization is gonna be really big

Training Day… fire movie

I kinda noticed that, lots of expats with strollers. But the city just felt strange to me - you can tell it was built up pretty recently and with the constraint that it was going to be extremely hot outside 99% of the time

I’m sure it’s purely a me thing/skill issue but I wasn't really taken by Singapore. Probably my least favorite big city in Asia so far.

Special preview of my talk on scheme-rs coming up this thursday

Since lifetimes express a subtyping relationship I tend to view it as top, and in the T: context expresses the minimum lifetime allowed, which works okay with dyn +

Pretty interested to find out if adding the concept of known functions improves perf in scheme-rs at all

I don't think it'll be hard to get decent performance in the areas scheme-rs is particularly week, but getting really amazing perf is going to be a long term issue

Desperately trying to optimize scheme-rs now that I set up a decent benchmarking framework to compare against other implementations. On the plus side, the fibonacci benchmarks are really good! On the minus side, everything else is pretty bad 😬

I think I may have developed a version of concurrent cycle collection that removes the mutation buffer entirely, which ends up being muuuch faster

Hahahaha “They went after me and Theranos this hard
because it was all make believe” is SO much easier to believe lmfao

my girlfriend and I found a Tomato plant on the sidewalk someone was giving away for free at the beginning of summer, so we decided to plant it. It never bore any fruit but it did get pretty big

probably the best shader of the last ten years IMO

Same. It also just look soooo cool. Hard to justify buying a new computer though

IMO there’s literally nothing better than a big pot of greens and beans