@ChShersh I think that would be std::weak_ptr. You think you own it, until the government takes it away from you.
English
Christian Hofer
901 posts
Christian Hofer
@chmhofer
Father of three. Gemini at work, Kimi at home.
Seville, Spain Katılım Aralık 2012
413 Takip Edilen96 Takipçiler
In the age of AI, knowing Vim still helps me daily.
Truly a timeless essential skill to have.
English
@ChShersh You don't like C++? Don't let the functional programming expert hear this!
English
It’s hilarious how someone writes a bad post about C++ and each time three absolutely random people summon me replies telling me I NEED TO SEE THIS AND DISAGREE
English
@ChShersh Well, instead of learning Zig, I ended up vibe coding in Python. Learning Ukrainian is probably more honourable... Although, I imagine, for you it must be like for me learning Dutch.
English
3 months in. My plan for 2026 is completely fucked.
1. I’m not learning a new programming language. I’m learning Ukrainian instead.
2. I was rejected to speak at a C++ conference.
3. I failed at reading.
4. My secret goal didn’t happen. In fact, I even forgot what it was.
Well.
Dmitrii Kovanikov@ChShersh
I reflected on my achievements in 2025. It's time to plan for 2026. My aspirational goals for the upcoming year: 1. Learn a new programming language 2. Present at a C++ conference 3. Start reading consistently 4. One secret goal I hope to unveil in 2 months Nothing truly ambitious. My life has changed dramatically in 2025. I never could've imagined the turn it would take. Jan 2025 looked completely different from Dec 2025. Both in terms of personal growth and my dreams. I struggle a bit with seeing a vector for my future. I tried many things in life. I want to try more, but I know it's not the right time or they don't align with my values atm. I demolished my past self. It's time to build upon a new foundation. Let's see how it goes. I hope you have something to look forward to. Happy New Year, everyone!
English
Dijkstra's algorithm was invented in 1956.
This theoretical paper was published in 2025.
The real algorithm that beats them both on real data AND PEOPLE ACTUALLY USE NOWADAYS was published in 2010.
Tech with Mak@techNmak
For 38 years, computer scientists believed Dijkstra's algorithm was optimal for sparse graphs. The logic seemed airtight: Dijkstra sorts vertices by distance. Sorting has a lower bound of O(n log n). Therefore shortest paths can't be faster. 5 researchers proved the assumption wrong. The trick: combine Dijkstra's priority queue with Bellman-Ford's dynamic programming. Divide and conquer on vertex sets. Shrink the frontier. Result: O(m log^(2/3) n) First improvement for directed graphs since Fibonacci heap in 1987. Tsinghua. Stanford. Max Planck. 17 pages.
English
@ChShersh For me that was quite a shock initially to learn that these "optimal algorithms" are not the ones that are useful in practice. Our algorithms department at university was focusing on the latter, and I was stupid enough to be disappointed by that...
English
@ChShersh Maybe I would be willing to trade your 2% of C++ against my -2% of C#, though.
English
This is the ideal tree bottom-up traversal. You may not like it, but this is what peak FP looks like.
Dmitrii Kovanikov@ChShersh
I'm non-ironically convinced this is the best representation of a tree in C++
English
Thanks, @lisyarus, for these examples of variadic templates.
They enabled me to finally implement a nice eDSL in C++ for creating TUI layouts declaratively.
Nikita Lisitsa@lisyarus
@ChShersh Yep, I mean scope-based automatic destructors specifically, aka RAII. As for variadic templates, I use them in a lot of places! My ECS implementation (bitbucket.org/lisyarus/pseme…) is heavily based on them, e.g. here's how I create an entity - world.create() has to use variadics
English
@debasishg Are they different to C# structs that have been around from the beginning?
English
value objects in JDK 26 will be a game changer in data oriented programming ..
Value objects are identity-less and immutable by design.
The JVM is allowed to:
• Flatten (inline) them directly into the memory layout of the containing object or array.
• Eliminate the per-instance header and padding that regular objects require.
• Store them contiguously, just like a C struct or a Rust struct.
Concrete performance wins delivered:
• Cache-friendly data structures: Point[], Complex[], Vector3d[], record-like data carriers, etc., now behave like primitive arrays.
• No more @ Contended hacks (the old annotation that manually padded fields to avoid false sharing).
• Better locality in collections (e.g., ArrayList or specialized value collections will be dramatically faster).
• Reduced GC pressure because fewer objects = fewer things for the garbage collector to scan.
• AOT / JIT friendliness - the compiler can make much more aggressive assumptions about layout and inlining.
Paul@PaulLarionov
JDK 26: - build-in HTTP/3 - AOT caching for lightning-fast startup - project Valhalla: value objects, goodbye memory overhead, hello CPU cache - project Loom: structured concurrency making async code as easy as sync - project Amber: pattern matching & data-oriented programming - project Babylon: running pure Java natively on GPUs for AI
English
@ChShersh @HaoTurnip How many GitHub repos are leveraging it? LLMs use constructs most C++ porgrammers would use.
English
@HaoTurnip It's crazy how try_emplace was added in C++17, yet LLM still doesn't leverage it 9 years later
English
Same C++ function.
One is generated with AI.
The other one is written manually.
Guess which one is which.
English
@stagerbn Was hängt schon wieder fest, "der Wal" oder "die Wahl"?
Deutsch
Der Wahl in seinem Tun steht exemplarisch für ganz Deutschland.
Deutsch
A timely measure to prevent our cosy programming community to be destroyed by all those people of Eastern European origin with their bizarre love for low-level languages! [Irony off]
Nikita Bier@nikitabier
Starting Thursday, we'll be updating our revenue sharing incentives to better reward the content we want on X: We will be giving more weight to impressions from your home region—to encourage content that resonates with people in your country, in neighboring countries and people who speak your language. While we appreciate everyone's opinion on American politics, we hope this will disincentivize gaming the attention of US or Japanese accounts and instead, drive diverse conversations on the platform. We invite creators to start building an audience locally. X will be a much richer community when there's relevant posts for people in all parts of the world.
English
The people who clearly believe this most are...
Men who are 5'11"
Financial Times@FT
If there is a status attached to height – and there is – then 5ft 11in doesn’t quite measure up. Six foot is manly, commanding. Five-11 is acceptable, unremarkable. But a near-miss. It’s the upper end of almost, the FT's Alex Bilmes writes. ft.trib.al/Cy2ronS
English
@ChShersh Sadly it's 2026 and line 1 still doesn't read:
import std;
English
The C++ example is wrong btw.
This is C++ in 2026.
Python Programming@PythonPr
Which language do you think is best?😍😍😍
English
@ChShersh You still didn't get over your Haskell past, did you?
English
Your C++ code without pointers — pointless.
My C++ code without pointers — pointfree.
We're not the same.
Dmitrii Kovanikov@ChShersh
Pointers in C++
English
@jdegoes Do you have a link? For me, it makes sense that Rust should work well. In my experience, Python is going much smoother, though.
English
@chmhofer Benchmarks don't support that hypothesis for certain choices of [my favorite language].
English
Bad take.
It's trivial to fully prevent coding agents from ever using unsafe--even, with the right harness, WHILE they are in the act of trying to use it!
Rust beats Go with stronger static typing and no GC.
In the agentic era, Rust > Go, and it's not even close.
Emil Privér@emil_priver
I don't think Rust is the best programming language for LLMs. The only reason why Greg said this is because Rust limits the LLM from making mistakes by not compiling shitty code, but the LLMs still suggest using unsafe! to bypass security rules. Therefore, I would say that Go is the best — if there is any "best" language for LLMs. Since LLMs are based on next-token prediction, it's more important that the code looks almost the same to get a good result. Go was designed to make every app around the globe look almost the same; for instance, there is only one way to make a loop compared to Rust, which has at least five that I could think of (loop, for, while, while let, and map). I want the best language to be rust tho
English