วีดีโอ

Thank you FPSRussia

How do i run the query? do i need to create a keymap or am i missing something?

Thanks for the explanations. I'm writing a tiny RTOS for the MEGA 2560 (that's because it's what I have) and what is propose by RIOS is close to what I'm looking but lack of small features. PORTB = 0xff is to enable all 8 ports of port B as outputs. 🙂

I WAS searching for this, and thought there's no tutorial for it on youtube and can't find a clear instruction either.. so i figured it myself. But now that I already know what to do, I finally found this video...

btw did you notice gopls finally has type inference for generics

Great video! Hey I'm wondering if you've halted development on dbee.nvim or if it's something that you'll jump back into at some point? Love the plugin and your videos, keep doing what you do!

Having Data Analysis, Data Science, AI in Python as the number of language, and then the Speed HAVE TO be a MUST for python these days.

So it’s basically swift with a new compiler ?

I wish json were more performant, it’s not hard to improve substantially. There’s a workaround for every deficiency in stdlib through a third-party package. But having stdlib on par with them out of the box in a consistent way would be really welcome.

Great video, subbed

The error handling in Go is fine I think. Some syntactic sugar may be nice, but not needed and makes the language slightly more complicated. Unions are still useful for many other cases though. What I would really like is to get away from runtime reflection and make compile time reflection the norm. Performance is really an important topic in more and more projects. Go generate is basically that but inside the tooling instead of the language itself. The tool directive shows they want to keep going into that direction. Fine for me. The rework of the JSON API is interesting. More performance and better behavior is great. Protobuf changes look good too, even though I haven't used it yet.

Glorious evolution!

watch?v=s7WNFzm3vjo

Subbed... thank you! (get better, kids!)

Subbed

As a recovering C++ developer, I *love* Go's error handling. I do not think it is broken and therefore needs no fixing.

This started with Python 3.11

we need a new c++ . from scratch

Hey nice plugin. In the example you showed, it uses docker to enter the database username and password. How do i enter username and password for a remote database not based on Docker?

very underrated

Typical C/C++ development th-cam.com/video/szcQjan875A/w-d-xo.html

Im probably being dumb, but what button do you press to execute the query? learning nvim and i just can't get the query to execute

I have seen many developers think that the infrastructure diagram is THE architecture diagram and do infrastructure driven development. I try to teach developers that having one runtime gives us enormous flexibility in the future, since we can run it locally, serverless, in docker containers without any changes. At my last project we had a monolith with one run time deployed as 5 different services with different autoscaling. All interservice calls could be performed in direct call mode, in http mode or in event mode, so our automated testing could be run as a monolith and we did full BDD testing, where most of our tests was end-2-end like.

MLIR is more about lowering a language in refined steps, and doing more optimizations and analysis on the high level representations. With LLVM, it's analysis passes had to try to identify high level language constructs and optimize them /through/ the low level IR, whereas MLIR offers a framework to do these optimizations and analysis on high level constructs directly

Not implementing concurrecy from the start might become problematic, since adding concurrecy to an existing codebase can be a real nightmare

I did a similar experiment as you did when 3.13 was released. I used the concurrent.futures.Executor abstraction so I could easily switch between ProcessPoolExecutor and ThreadPoolExecutor via command line switches. I used a brute force search for Perfect Numbers up to some value of N (also passed on the command line). I found the results somewhat predictable, but with one surprise. With the standard python binary, the single-thread / single-process (naive) implementation was over 15x slower than multi-process (10 workers). The multi-thread (also 10 workers) was, no surprise, at par with the naive implementation. This clearly shows the effect of the GIL. The results with the 3.13t experimental NoGIL binary was about the same for the multi-process impl. The naive impl was slower(~20% slower than the standard python binary). But the real surprise was that the multi-threaded impl was actually slightly slower than the multi-process impl with either binary! Of course, this shows that there is more context switching overhead with threads in a single-process vs multiple kernel-managed processes. And, of course, there is probably some noise in my results due to the architecture of the laptop on which I ran the tests, but the results should be comparable. I concluded that it is best to just stick with the multi-process parallelism approach for now. At least until further experimentation with NoGIL leads to some further Python improvements. I have the experiment posted in a github repo. Let me know if there is any interest in my posting the link. Thanks for the great video! I really appreciate seeing someone else sharing roughly the same results that I experienced.

We have mojo for speed

c++ is in fact the BestLanguageOfAllTime because its c++

is this going to be a experimental feature or in the future will it be default?

We don't have to agree in the details, and I take that is your point. In the end, use your head is the only guide we shall stick to. Good video. I agree completely with this after almost 30 years in the software industry. There's no free lunch.

Splitting services into specific, smaller "microwave," "fryer," and "stove" microservices makes each one simpler and faster to understand. Onboarding is easier, and if needed, reimplementing just the "microwave" service is far less painful than tackling a monolith-service. Plus, with clear gRPC proto definitions, you get built-in documentation, so you don’t need deep dives into inner workings-unlike larger services that grow in complexity and need detailed UML diagrams or process charts to explain.

sorry i don't know what is a micro service all my backends are one 300000 line repo

CDD will get you fired because you can't implement features fast enough. Yes Micro services are terrible but CDD is not he solution

true

Awesome video. The use of microservices is an optimization, and almost always a premature one at that. Start from a monolith, then split off microservices as you discover a need.

I worked a lot with Microservice like projects in the context of real life enterprise applications, in my opinion, you are right on a lot of points, and what I agree with you is that Microservice Architecture is not the Go To Solution for any project, yes sure it depends. I witnessed a project where adopting microservices make it an Endless Spaghetti Nightmare and I saw another legacy project a Gegalith that the build takes 2h+ with constant application outage and very poor performance ( a business report generated in 20m+ if it didn't fail). In the end, change is constant we have to know how to deal with it in an elegant Agile way and Software Architecture is an Art, the patterns are there for lessons.

So it's a proposal where C++ gets another language embedded into it. You of course have to opt in to use it and it's probably going to be a billion new keywords and syntax. Sounds familiar doesn't it. The problem with C++ is the fact that it's too overbloated. If you need the safety features, you really don't have an option but use Rust.

Thats why Nim is the best language.

real sigmas leak memory

The state after removing apple should contain 3 keys (grapes, banana, orange), not 2 as visible in the video, so it is in fact not correct

Yup this was really fun. I recently just learned about Hash tables too so this was really nice to see. Thanks boss!

Great plugin. Thank you o/

it feels kinda the same as python. it's just not much different for what I'm doing

Zig error handling would be great

With modern C++ practices one hardly encounter memory safety issues. When you use it in a C like way you run into above issues.

"Rust Lobby" sent me lmao the meds mason what do they say?

outdated

Lua/LuaJIT >>>>>>>

Waiting for Mojo to get some more mature.

llvm - architecture mess, thx for vendors its living swift - is barely working transpiler to C++ mojo - ?? hmmm worlds slowest statically typed language? 1:25 mlir is focused on parallelisation