4,199 karma · joined December 10, 2014
I think one of the clearest examples would be drivers -- they might be properly sandboxed by a good microkernel, but they are notoriously buggy/crashy/incorrect. If I remember correctly one of the Fuchsia team members told me that they were going to support drivers written in Rust, but I could be completely wrong.
Filesystems, the network stack, whatever horrifying systemd equivalent Fuchsia grows, etc. are all examples of things I would advocate for writing in Rust before focusing on the kernel. All of these things are still security/reliability sensitive, all still terrible by Rustacean safety standards in most OSes, and conveniently don't have to be built directly as part of the kernel when you're doing a microkernel.
Just spitballing, of course.
Uh, this is rambling a bit. My point is that building zircon as a microkernel makes it much easier to write correct code, Rust or not. And that maybe the overhead of a rewrite wouldn't be as beneficial for the microkernel as it would be for other parts of the operating system.
We just use this protocol/schema to make sure that development URL links are tappable on your phone. Checking a URL domain works for lots of deep links, but not when you have IP addresses or dynamically generated ngrok URLs (like when loading an app from XDE).
You might ask, why do this when you can just scan or screenshot the QR code? Fun bit of tech debt: the current QR-scanning system was added to support the offline-only create-react-native-app setup, and the "send a link by SMS" flow predates that. If you've received an SMS with your development URL, it's pretty useful to have it tappable.
Related, this detail of the XDE <-> client app connection isn't really what people use for production, it's just for development and sharing with coworkers in most cases. To ship a "production" Expo app, users typically build an IPA or APK for uploading to the App Store and/or Play Store, where this URL is hidden from them.
Note in that thread that the author implies that the optimization levels between the two tools are not necessarily equivalent or stable.
$ optipng --version
OptiPNG version 0.7.6
$ oxipng --version
oxipng 0.16.3
Ran on a 2016 MBP, i7.Time:
opt-level oxipng optipng
2 148.63s 30.80s
3 9.56s 45.99s
4 23.82s 90.57s
Size of resulting image: opt-level
original 12M
oxipng optipng
2 12M 12M
3 12M "already optimized"
4 12M "already optimized"
TBH I don't really know how to interpret these results. But with the strange exception of `oxipng -o 2`, it does seem like the naive parallelism offers some real performance wins. Disclaimer: this is an incredibly amateur attempt to benchmark, I really don't have enough knowledge to produce something authoritative.If others step in to offer support contracts as well, great! Competitive environments for that would probably be good.
Presumably your consulting business has really taken off now that you have such unique insight into a historically unempirical area of our field?
Assuming they have something like rainbow tables for short passwords on all of these managers, it still seems like it would take a very long time to correctly guess longer master passwords (say 20+ characters). No?
I've interviewed candidates from them - they're also really great for this.
"That’s because I have been working on this subject for a long time now:
* I did multiple talks on it * I even co-wrote a paper * I did it both as client and personal work"