_vvhw

42 ms·

151.
by _vvhw 5y ago
If I "have utterly failed to understand" destructors then I'll wear it gladly. :) I suppose we differ mostly not in our view of destructors, but in our approach to reducing dimensionality and how best to do that.
152.
by _vvhw 5y ago
You're conflating undefined behavior with spatial/temporal memory safety, the latter of which is what is 100% prevented by Rust's borrow checker, provided you're not interfacing with hardware, in which case I believe tha
153.
by _vvhw 5y ago
Also Zig's memory alignment in the type system is great for doing low-level I/O. Coming from C, I find that Zig's memory alignment options are easier and more powerful.
154.
by _vvhw 5y ago
> But he is right that "no hidden control flow" is an anti-feature. Hmmm, "no hidden control flow" is at least what I want writing distributed systems, storage engines or device drivers. Calling it an "anti-featu
155.
by _vvhw 5y ago
> "read the source code" comments always strike me as unhelpful. I can understand why people would recommend reading the source code of projects like Redis or SQLite (and here Zig's std lib) to learn, simply because the co
156.
TigerBeetle's LSM-Forest(youtube.com)
3 points by _vvhw 5y ago|0 comments
157.
Assembly Nights(ratfactor.com)
123 points by _vvhw 5y ago|36 comments
158.
Viewstamped Replication Revisited(charap.co)
20 points by _vvhw 5y ago|0 comments
159.
by _vvhw 5y ago
By the way, thanks Thomas... I've always enjoyed and benefitted from your "CSPRNGs are essentially a stream cipher, and now you finally also understand what a stream cipher really is" explanations on HN over the years. I reme
160.
Do We Need a Link Step?(robert.ocallahan.org)
2 points by _vvhw 5y ago|0 comments
161.
by _vvhw 5y ago
Ah yes, good point! I was assuming in both cases of manual failover that the operator would have to have some way of physically shutting down the old primary, then starting it again only as a backup that doesn't reply to clients. Alter
162.
by _vvhw 5y ago
What are the distributed options for Postgres? What mechanisms are available to make it highly available i.e. with a distributed consensus protocol for strict serializability when failing over the primary? How do people typically deploy Pos
163.
by _vvhw 5y ago
Agreed, but I think that many people would consider Unicode normalization to be part of what they want from the std lib when they mean that UTF8 should be baked in... so that they can manipulate UTF8 as they want, including in various norma
164.
Viewstamped Replication Revisited(charap.co)
2 points by _vvhw 5y ago|0 comments
165.
by _vvhw 5y ago
Also excluding Unicode normalization? Or should that also be baked in?
166.
by _vvhw 5y ago
I love the idea about compiler checking for unused variables in production, and that it's strict and not a warning that can be ignored merrily merrily, gently down the stream! :) We found a few cases in TigerBeetle where some nasty bug
167.
by _vvhw 5y ago
Zig's a little different here. Zig's C interop means it can leverage some of the most battle-tested libraries out there, with no shortage of libraries given C's massive ecosystem. And with Zig, it's not like you have to
168.
by _vvhw 5y ago
Zig's C interop is pretty good though, and there must be some decent native Unicode library out there somewhere right? ;) I've worked on a full-duplex file synchronization system that had to support cross-platform operating system
169.
by _vvhw 5y ago
And where the CPU doesn't support some SIMD intrinsics—SWAR for the win! I've been digging into Zig's Vectors the past week and really like the decisions. Thanks!
170.
by _vvhw 5y ago
Which version of UTF8?
171.
by _vvhw 5y ago
No, I believe it's the right decision and I respect it. Encoding charsets are continuously evolving and shouldn't be baked into a language specification.
172.
by _vvhw 5y ago
I loved your talk on this ("Zig in Production"): https://www.youtube.com/watch?v=124wdTckHNY
173.
by _vvhw 5y ago
My guess is that Zig's design choices are hitting "a sweeter sweet spot" for systems programming that resonates with many engineers reading HN. At least for TigerBeetle [1], a distributed database, the story was strong enough
174.
by _vvhw 5y ago
"And just because the source filename has to end in .gif, that doesn't mean it's really a GIF file. The ImageIO library, as detailed in a previous Project Zero blogpost, is used to guess the correct format of the source file
175.
Hytradboi(hytradboi.com)
4 points by _vvhw 5y ago|0 comments
176.
by _vvhw 5y ago
With context switches becoming more and more expensive relative to faster and faster I/O devices, almost the same order of magnitude, I believe that thread-per-core is where things are heading, because the alternative of not doing thre
177.
Assembly Nights(ratfactor.com)
1 points by _vvhw 5y ago|0 comments
178.
Connecting the Internet Economy(fynbos.dev)
3 points by _vvhw 5y ago|0 comments
179.
River, a Dynamic Tiling Wayland Compositor(isaacfreund.com)
4 points by _vvhw 5y ago|0 comments
180.
by _vvhw 5y ago
I like the notation: A(-:-,100) B(100:b,100) C(100:b,-) Just reading to that line, it was clear: what was going to go boom, and what the solution(s) would be.

More ›