olix0r.net
Creator, Linkerd š¦
Co-founder, Buoyant š
Rust š¦, Async š, Networking šøļø
65 posts
1,914 followers
93 following
Prolific Poster
Conversation Starter
comment in response to
post
There's a theory that Hangzhou's business model is to short nvidia and ms... seems like a stretch, but it's a fun one.
comment in response to
post
Thatās reasonable! But down the rabbit hole we goā¦ are you enforcing guaranteed QoS? If so, how do you figure out what your workload needs so that a safe limit can be set? If not, do you understand how your system behaves when the node is oversubscribed?
You get my pointā¦ not trivial!
comment in response to
post
comment in response to
post
Neat!
comment in response to
post
Iām always thrown by āInternal Compiler Errorā
comment in response to
post
Yeah.
comment in response to
post
Itās 1.0 and part of tokioās public api. Itās dependable.
comment in response to
post
Would make futures 0.3 part of the public api surface area.
comment in response to
post
And the most effective solution is controlled burns! but communities hate them and fight them tooth and nail.
comment in response to
post
honestly i try to draw inspiration from tokio whenever possible. some of the best organized library code out there imo.
comment in response to
post
yeah. here's a maybe overkillish example, including feature flagging: github.com/olix0r/kuber...
comment in response to
post
I usually see public reexports at the lib.rs level, with the inner module structure largely hidden. Also note judicious use of non_exhaustive on public structs for api stability
comment in response to
post
Remember when we were all concerned about the power telcos wielded after the patriot act? A lot of great tech came out of that.
comment in response to
post
shouldn't they call it "fucking sign on" then?
comment in response to
post
are you new here?!
comment in response to
post
comment in response to
post
this is rad!