eidolon @shaderphantom@awoo.space

oh yeah. no upcasting. that's why.

maybe it's because trait objects are constructed very differently from C++ vtables that we tend to shy away from them. but maybe trait objects really could help us solve a broad class of issues without diverting to more academic data-oriented design solutions and without losing out on that much performance

this leaves me to question the Rust ecosystem's approach of making everything generic by default; monomorphization is expensive though the runtime is very very fast, but maybe the value in monomorphization isn't universal and we should be using trait objects more than we do?

huh. AActors also are just associated with root USceneComponents in the world, which is why actors aren't treated like a scene graph.

which leaves the question, how does the engine handle multithreaded anything with the way it's architected? for as modern as this engine is, it seems very traditional in its architecture...

fascinated that UE4, the world class game engine known for high performance rendering, seems to use virtual functions everywhere even on the hot paths.

Hey fuck the web actually

this use case may be better suited to Cassandra but that's not what we have today right now

worse yet, this is Amazon Aurora MySQL -- shouldn't their storage driver be better than this considering the pricing?

I'm really confused because this table doesn't even have any foreign keys or external references, only a handful of columns, and the primary key is just an incrementing integer. this is just a normal use case, but inserting 200 rows via batch insert takes upwards of 4 or 5 seconds. I purged the table and did the same insert and it still took about 700ms.

hey does anyone know why mariadb/mysql would chug hard when inserting rows into a table with only 1 million rows and a single secondary index on a single column

chrome's near monopoly is worse than the ie6 nightmare

it's actually just broken: the CSP the page serves includes directives that override the unsafe-inline declaration anyway. somehow chrome thinks this is normal and goes along with it, but firefox rightfully errs on the safe side?

of all the industries trying to modernize for the internet still, i think the healthcare industry pulls the most nonsensically insecure and dangerous bullshit

this is because their login page is an embedded okta login which requires unsafe-inline enabled in the content security policy (!!!! what the fuck!!!!!), and firefox seems to reject this by default

a certain major healthcare identity provider for a certain major hospital network has managed to completely break their login page on firefox

hey folks, haymarket books is running a 50% sale on books about fighting fascism

https://www.haymarketbooks.org/blogs/81-50-off-haymarket-books-for-fighting-fascism

incredible resting action

wow! snoozin is good

want to sleep!