Tokyo station entrance ceiling. More impressive in person.

Surprisingly this is potato salad and not ice cream. Still delicious.

I recently went out to a Mexican izakaya. If you're curious about Japan's take on Mexican food, you might find it interesting: byuu.page/2019/11/mexican-izak

I've been taking a bit of a breather on higan and bsnes development for a bit. Still poking in to fix any critical issues and merging patches, of course. But it's been a mad rush (an enjoyable one at that!) since around March of last year, so a bit of time off feels like a good idea.

I like how resistant Zelda 2 is to a remake. The game flat out won't be anything like the original as 3D, nor as a 2.5D style game. Combined with it being underrated, it means it can be left alone and enjoyed as is. That's rad.

@stsp @oa Casually name-dropping Theo: Buehler or de Raadt? The latter would be quite intimidating to have look at my evil cooperative threading hacks, hahah.

@stsp @oa Hell, one last problem: libco now has co_derive to use user-supplied memory, which is meant to accelerate thread recreations without having to constantly malloc/free, and new higan/bsnes builds it. This is just overall going to be a disaster ._.

@stsp @oa Second random thought: write the stack size into the first 4 or 8 bytes of the allocated stack memory pool if you really need it, and then extract it when deleting the thread. It's still a pain if I have to do that for every OpenBSD target though.
What are you even doing with MAP_STACK? Do Intel CPUs even support throwing an exception when the stack pointer is not in a specified range?

@stsp @oa Alternate random thought: what if you hack libco.c to use the SJLJ sigaltstack trick for OpenBSD? It's not as performant but you guys care as much about security as I care about accuracy so it should be fine ^-^

Show more
Mastodon

Server run by the main developers of the project 🐘 It is not focused on any particular niche interest - everyone is welcome as long as you follow our code of conduct!