Proposal to make programming more appealing to Gen Z by renaming keywords typealias -> typeshit

The fourth move: tag each kind by domain. `User.Count` and `Inbox.Count` are both `UInt`-backed counts, but they won't add — different tags. Write the typealias once per domain, the compiler keeps them apart from there. Same shape for `Element.Index`, `Slot.Index`, etc.

typealias EverythingDoable = Codable & Sendable & Hashable

typealias addr_v8_t = uint64_t;

Have you tried name-based destructuring in Kotlin 2.3.20? In this 8-minute clip from the recent TypeAlias Show Livestream (link below), you can learn all about the three modes of this new experimental feature, and see how you can start using it (or migrating toward it!) in your own projects. youtube.com/watch?v=tvpB6_6x…

Finding 10 | Undocumented API features in proto definitions vs docs.x.ai EFFORT_MEDIUM reasoning level: grep -n "EFFORT_" xai-proto/proto/xai/api/v1/chat.proto Output, 540: EFFORT_LOW = 1; 541: EFFORT_MEDIUM = 2; 542: EFFORT_HIGH = 3; Proto line 140: Default to EFFORT_MEDIUM. SDK type definition: grep -n "ReasoningEffort" xai-sdk-python/src/xai_sdk/types/chat.py Output: ReasoningEffort: TypeAlias = Literal["low", "high"] The medium level is defined in the proto with a comment indicating it is the default, but it is not exposed in the SDK type alias.

Kotlin 2.3.20, Android Bench, and More - TypeAlias Show #10 x.com/i/broadcasts/1nJOLEeQR…

Yep… annoying… But my trick is to extend the generic structure with <Any> and typealias it as AnyTimeDetailView, then put those enums in there.

These two typing features look similar... But they're very different. Here's how to use NewType and TypeAlias in Python 👇

Sometimes, I manage to write code that makes me genuinely proud! 🌟 The entire service is built on a simple abstraction, utilizing generic types and depending on a specific typealias. ✨ It works beautifully and looks stunning! 🤩 #BuildInPublic #DevDiary #Permissions #Swift

why `type` statement is better than `TypeAlias`

"public typealias" "try await" estou encantado

ちなみに Swift 2.1 までは関連型の宣言も typealias でした github.com/swiftlang/swift-e…

associatedtypeはプロトコルのパラメータであるメンバータイプ要件で、メンバーメソッド要件などの仲間 typealiasは型の定義方法の一種で別名を与える

associatedtype typealias 違い

back to using Rust again after 1.5y, and having a really really great time using it with Claude Code. personally beats Bun for scripting too? Notes on why: - context: my main task rn is to write various complex scripts which munge data from websites & GCS, usually writing them back to GCS - at reasonably large scale: I’m def trying to parallelize & unbottleneck a lot - I used Python & Bun each for a bit, but maybe prefer Rust? - There’s a reasonable amount of vibe coding. I know the precise steps quite well but I’m often asking it to eg: list X kind of file from GCS, parallelize Y, fetch Z in ways Claude is good at. - typing is so much better! - makes me reflect on the fact that actually most type systems are expressive enough that it doesnt matter. - the real important factor is level of typing in the ecosystem, esp for small projects where you interface w many libraries. - ie: Python typing is bad because many many libraries lack it. Even with TS, major libraries will sometimes still typealias any - errors being passed as optionals is really nice. Again in TS, you could do this via effect TS, but your upstream libs won’t really? - you’re also likely writing out of distribution TS which models are likely worse at - way fewer logic errors in general! Cargo check is good because typing is more prevalent. - I totally forgot about this, but it often feels like the Rust libraries are just better? - the default ish CLI arg parsing library feels so much more pleasant than argparse or commander. It’s very typey - for example, frequent operation is, I’m downloading 1k files. Do this in parallel w 50 at a time. - Claude writing a Bun script will always do the stupid thing of slicing into batches of 50, and sequentially promise.alling each batch. This is much slower, because you’re seldom actually at 50. Particularly extremely bad if some tasks can take >10x avg time - the right approach is to remind Claude to use something like p-limit. It sometimes still gets syntax wrong? - otoh: in rust it just does the reasonable thing 100% of the time because a semaphore is a common primitive - in general, it feels there’s almost this inductive bias causing models to write better Rust than they do TS? - the default Rust way to do things is more likely the right way than a hacky way. - the model basically ~never coerces types or does weird typing stuff - the default first library for something is on average really good

#Kotlin Efficiency: Code Smarter, Not Harder - Typealias #dormosheio #programming #androiddev #mobiledev #iosdev dormoshe.io/trending-news/ko…

🎙️ Live this Wednesday — From Koin to Kotzilla! We’re joining @djleeds on the TypeAlias Show with @arnogiu and Miguel. We'll debunk @insertkoin_io myths, share updates, and talk performance improvements. 📅 Oct 15, 5PM CEST 🔗linkedin.com/events/73835584… #AndroidDev #kotlin

Swift の TimeInterval は Double の typealias なので、0.1 0.2 = 0.333333... の問題が起きうります。 Duration は typealias 等ではなく、これも考慮されているはずなので起きなかったはず… です……。 #iosdc

Finally fixed a strange link issue of private AG API. It turns out to be kind=Structure vs kind=TypeAlias for imported clang module mangling rule. See detail in github.com/OpenSwiftUIProjec…