There’s a setting that controls whether it will do a snapshot first. Turn it off and it will just start sending through new cdc entries.
> you must set the kafka retention time to infinity
Is this a new retirement? I’ve never had to do this.
7,095 karma · joined March 20, 2017
There’s a setting that controls whether it will do a snapshot first. Turn it off and it will just start sending through new cdc entries.
> you must set the kafka retention time to infinity
Is this a new retirement? I’ve never had to do this.
That is a _technical solution_ to a fundamentally _social_ problem.
Very much of the opinion that the Unikernel stuff (and especially what UniKraft are offering) is being massively slept on.
I’ve learnt things and been exposed to ideas developing software for work that I simply wouldn’t have if I was only doing it in my spare time.
Mmhhmm, yes I’m sure you Know Better™ but maybe don’t complain when you do this to yourself, and then it makes your life hard?
> If your optimization is that situation specific, it should stay stuck in "experimental".
Yeah, I have a feeling that there’s low-key lots of applications for whom many compiler optimisations that you’re branding “experimental” are in fact run-of-the-mill and would just enable “experimental” mode so frequently it’d just be synonymous with “actually release mode”.
> And through all of this, the compiler should also eject files that carry enough information to allow debuggers to make sense of the compiled code, unwind the optimizations when the users asks, and present a coherent version of what is going on.
This is a pretty cool idea, semi-related, you should check out some of the work around E-Graphs, which attempt to enable the same level of optimisations but also preserve more original context. It’d be neat to have the equivalent of source-maps for highly optimised code, but I’m not sure how we’d manage them without them becoming enormous. After things like inlining, constant-folding, desugaring, rearrangement operations, etc all take place I suspect you’d be left with code that bears only the merest passing resemblance to its original form.
> Could people read terms of service…
Even if they do read licences and such, companies have a vested interest in making them as complicated, obtuse and self-serving that you have close to no recourse. It’s weasel-worded to the nth degree. They also change them largely at their leisure, and if the new terms are bad, again, there’s often very little you can do.
“If consumers don’t like it, they wouldn’t buy it” is the other lie that’s successfully kept itself alive. Consumers are kept time and spare-resource poor, and are largely presented with a predefined set of options to choose from that the companies at play feel like presenting us with. Rarely is there an _actual_ varied choice. Only the illusion. Combine that with scenarios in other industries like enterprise sales where the “customer” is an exec and the user just gets lumped with some garbage software.
I have a friend who works in the networking division of a telco in my country, their team had to spend significant time and effort educating a PM who was dead-to-rights convinced that IPv6 was “less secure” and seemed to think that IPv6 didn’t have subnets and that NAT’s were the same as firewalls and refused to be convinced otherwise.
People like that make any forward progress extremely difficult.
NPM has done a pretty great job of showing everyone else what to avoid doing.
The mere mention of “web pack” sends most of the FE devs I’ve met into borderline trauma flashbacks.
There’s seemingly half a dozen package managements tools, some of which also seem to be compilers? There’s also bundlers, but again some of these seem integrated. Half of the frameworks seem to ship their own tools?
HTTP/3 is HTTP over QUIC. HTTP protocols v2 and onwards use binary headers. QUIC, by design, does 0-RTT handshakes.
> Yes, in theory UDP is faster than TCP but only when you completely abandon integrity
The point of QUIC, is that it enables application/userspace level reconstruction with UDP levels of performance. There’s no integrity being abandoned here: packets are free to arrive out of order, across independent sub-streams, and the protocol machinery puts them back together. QUIC also supports full bidirectional streams, so HTTP/3 also benefits from this directly. QUIC/HTTP3 also supports multiple streams per client with backpressure per substream.
Web-sockets are a pretty limited special case, built on-top of HTTP and TCP. You literally form the http connection and then upgrade it to web-sockets, it’s still TCP underneath.
Tl;Dr: your gripes are legitimate, but they refer to HTTP/1.1 at most, QUIC and HTTP/3 are far more sophisticated and performant protocols.
Many, many people I’ve dealt with in these roles don’t or can’t, and seem to think their sole task is to mainline customer needs into dev teams. The PM’s I’ve had who _actually_ do manage back properly had happier dev teams, and ultimately happier clients, it’s not a mystery, but for some reason it’s a rare skill.
Oh man. Once had a founder who did this to the dev team: blurry, pixelated screenshots with 2 or 3 arrows and vague “do something like <massively under specified statement>”.
The team _requested_ that we have a bit more detail and clarity in the designs, because it was causing us significant slowdown and we were told “be quiet, stop complaining, it’s a ‘team effort’ so you’re just as at fault too”.
Unsurprisingly, morale was low and all the good people left quickly.
Isn’t this a fairly well-known phenomenon though?
- new language makes waves. The people who picked it up early do some impressive stuff.
- early early adopter companies either snap them, or internally make the choice to learn it too. In turn, they do some stuff with it.
- gets a reputation for being the hot new thing. Other companies “want in on it”- they want to be able to do the fancy cool things the other places did, but they don’t have the patience, time or culture to grow it themselves so they aim to hire out seniors and everyone with lots of prior experience. <——- many orgs are here.
- proliferates enough that it’s well and truly mainstream. Also known as the “hire 1000 Java devs and throw bodies at stuff” stage of hiring and availability. Python, Node, Java, PHP, etc are here.
The Rust codebase I’ve built on run faster, don’t crash, are more correct, and take only marginally longer than the first python version. They’re easier to maintain, they’re easier to upgrade their dependencies, I’m more confident about asserting guarantees from the codebase, and it’s an easier time onboarding people and having them meaningfully contribute. There’s no more death-by-a-thousand-cuts that I experienced with Python codebases. I don’t have to worry about my Rust codebases blowing up at 2am because of some random exception.
I am defs not going back.
Can I hire 1000 Rust devs at the drop of a hat off the street? Well, no. Have I ever struggled to find competent Rust devs, or high level people who were happy to learn? Nope.
I would absolutely not be afraid to use it because of “hiring concerns”. If your current devs can’t or won’t learn anything new, they’re not very good at their job. If you can’t or won’t hire because you can’t find a super-senior at the drop of a hat, well, your hiring process is broken and no language will really help there.
Wanted to jump on the platform? Wow too bad, model had a little day-dream and now you’re floating somewhere. Wanted to peak a corner? Oops too bad, model had a moment and now the wall curves differently. Trying to pick a shot-well you’d definitely hit the other player if the model wasn’t making physics go out the window. Inconsistently.
We all thought Sports-Game $CURRENT_YEAR or CoD/MW MicroTransaction-Garbage was bottom of the barrel, a whole new world of fresh hell awaits us!
That is, seccomp is (apparently? I’ve never used it myself) capable of intercepting direct calls. Obviously, that design isn’t going to be able to handle “indirect” calls in its default implementation.
Either seccomp needs a way to act on the buffer or intercept io_uring calls, or there’s a need for a new auth mechanism that’s capable of handling io_uring style API’s.
Torpedoing the whole api (a la gcp) feels like throwing the baby out with the bath water.
> A plumber isn’t going to refuse adding another bathroom to your house, they’re going to tell you how it can be done and tell you the cost
Let me know how the plumber reacts when you then tell him said bathroom needs to be freestanding 300m in the air, and the plumbing needs to be made out painted twigs and twine (because your housemate thinks they look pretty), and you also want heated stone floors and you have a budget of $250 and an ice-cream wrapper.
In the miraculous case said plumber accepts the job, make sure to change specs halfway through.