If I were to give this post a longer title, it would be "I made a terminal pager because I needed a really good viewport component for my Go TUIs, then realized that a TUI viewport is just a mini terminal pager and I want the same text navigation and manipulation experience everywhere that I encounter long text blocks in my terminal".
I take no issue at all with `less`, it's super powerful and configurable as I call out in the post. I took the functionality I needed, made it a reusable component for Go TUIs, then made a terminal pager in the form of a Go TUI with it.
The reasonable suspicion around random people wearing secret face cameras is consistent with the idea that both more government and more corporate surveillance are bad.
Been fortunate to get to try out Sculptor in pre-release - it's great. Like claude code with task parallelism and history built in, all with a clean UI.
What exactly drove you nuts? The python ecosystem is very broad and useful, so it might be suitable for the application (if not, reasonable that you'd be frustrated). With strict mypy/pyright settings and an internal type-everything culture, Python feels statically typed IME.
It's not even close compared to working with Java or Go or any language built with static typing in mind.
To be clear, I'm not opposed to type hints. I use them everywhere, especially in function signatures. But the primary advantage to Python is speed (or at least perceived speed but that's a separate conversation). It is so popular specifically because you don't have to worry about type checking and can just move. Which is one of the many reasons it's great for prototypes and fucking terrible in production. You turn on strict type checking in a linter and all that goes away.
Worse, Python was not built with this workflow in mind. So with strict typing on, when types start to get complicated, you have to jump through all kinds of weird hoops to make the checker happy. When I'm writing code just to make a linter shut up something is seriously wrong.
Trying to ad typing to a dynamic language in my opinion is almost always a bad idea. Either do what Typescript did and write a language that compiles down to the dynamic one, or just leave it dynamic.
And if you want types just use a typed language. In a production setting, working with multiple developers, I would take literally almost any statically typed language over Python.
But TypeScript erases (its) types at runtime, exactly like Python. Python is Python's TypeScript. Whether you want TS or JS-like semantics is entirely dependent on whether you use a type checker and whether you consider its errors a build breaker.
I'm not sure what you're trying to say here. If you mean Python's type annotations are erased at runtime... Okay? It still has runtime type information. It's not "erasure" as that term applies to Java for example. And Typescript compiles down to JavaScript, so obviously it's runtime behavior is going to be the same as JavaScript.
In my view it's always a mistake to try and tac static typing on top of a dynamic one. I think TS's approach is better than Python's, but still not nearly as good as just using a statically typed language.
The fact that the types are reflected at runtime is what makes FastAPI/Pydantic possible, letting us use Python types to define data models used for serialization, validation, and generating. In TypeScript, we have to use something like Zod, instead of "normal" TypeScript types, because the types are not reflected at runtime.
I think a couple of things have to be untangled here.
The problem we are talking about in both Python and TS comes from the fact that they are (or compile down to) dynamic languages. These aren't issues in statically typed languages... because the code just won't compile it it's wrong and you don't have to worry about getting data from an untyped library.
I don't know a lot about Zod, but I believe the problem you are referring to is more about JavaScript then TS. JavaScript does a LOT of funky stuff at runtime, Python thank God actually enforces some sane type rules at runtime.
My point was not about how these two function at runtime. My point was that if you want to tac static typing onto a dynamic language, Typescripts approach is the better one, but even if can't fix the underlying issues with JS.
You could take a similar approach in Python. We could make a language called Tython, that is statically typed and then compiles down to Python. You eliminate an entire class of bugs at compile time, get a far more reliable experience then the current weirdness with gradual typing and linters, and you still get Pythons runtime type information to deal with things like interopt with existing Python code.
Typescript requires a compiler to produce valid Javascript. Python 3 shoved types into Python 3 without breaking backwards compatibility I think.
You would never have typing.TYPE_CHECKING to check if type checking is being done in TypeScript, for example, because type hints can't break Javascript code, something that can happen in Python when you have cyclic imports just to add types.
I would say mypy is better than nothing but it still misses things sometimes, and makes some signatures difficult or impossible to write. I use it anyway, but patched-on static typing (Erlang, Clojure, and Racket also have it) seems like a compromise from the get-go. I'd rather have the type system designed into the language.
I went from mypy to pyright to basedpyright and just started checking out pyrefly (the OP), and it's very promising. It's written in Rust so it's very efficient.
For the kind of work I'm using Python for (computer vision, ML), not really. The ecosystem isn't there and even when it's possible it would be much less productive for very little gain. Typed Python actually works quite well in my experience. We do use C++ for some hand-written things that need to be fast or use libraries like CGAL, but it has a lot of disadvantages like the lack of a REPL, slow compile times and bad error messages.
You have to force push each time you do this, right? How do your coworkers find the incremental change you made to commit 1 after you force push it, and how do you deal with collaborative branches effectively this way? And if I don't want to work this way and force push, are there other benefits of jj?
the heuristic is 'if you know about rerere and especially if you use it, you should try jj'. if you never force push, you might not see value in jj. (I basically always force push.)
JJ has the concept of "immutable changesets" -- if it sees a commit is referenced from a branch that it's not tracking, it assumes it ought not rebase that commit. Changesets on branches that look like the main branch are immutable too. And you can edit the revset that JJ considers immutable if you need it to be different from the default.
The net effect is that I can change "my" branches as I wish, but I can't change stuff that's been merged or other folks' branches unless I disable the safety features (either using `--ignore-immutable` or tracking the branch).
JJ also makes it really easy to push a single changeset as a branch, which means as you evolve that single commit you can keep the remote updated with your current work really easily. And it's got a specific `jj evolog` command to see how a specific changeset has evolved over time.
It's generally fine if you force push a branch that you're the only one working on. In many projects, there's an expectation that the 'PR Branch' you create in order to make a github pull request is owned by you, and can be rebased/edited/force-pushed at will. It's very common to do things like `git commit --amend --no-edit` to fix a typo or lint issue and then force push to update the last commit.
This has it's problems, and there's a reason things like Geritt are popular in some more sophisticated shops, as they make it much easier to review changes to PRs in response to reviews, as an example.
The PRs are either small enough that it isn’t a problem or large enough that it isn’t a problem… the odd in-between PR experience sucks and it’s one of the cases when I sometimes add more commits instead of force pushing.
+1 to sibling gerrit recommendation; I used to use it a decade ago and it was better then than GitHub PRs today.
This is great! What I would love is a way to compare myself with someone else though. I'm French and my wife is American, we have a lot of disagreement about colors (neither of us have vision deficiencies, we have ruled that out).
Somewhat similar to a site I made a while ago, but for more "perception boundary" colors: https://theleo.zone/colorcontroversy/
reply