o11c

The problem is that there's a severe hole in the ABCs: there is no distinction between "container whose elements are mutable" and "container whose elements and size are mutable".

(related, there's no distinction for supporting slice operations or not, e.g. deque)

True, speed does matter somewhat. But even if xterm isn't the ultimate in speed, it's pretty good. Starts up instantly (the benefit of no extraneous libraries); the worst question is if it's occasionally limited to the framerate for certain output patterns, and if there's a clog you can always minimize it for a moment.

Speed is far from the only thing that matters in terminal emulators though. Correctness is critical.

The only terminals in which I have any confidence of correctness are xterm and pangoterm. And I suppose technically the BEL-for-ST extension is incorrect even there, but we have to live with that and a workaround is available.

A lot of terminal emulators end up hard-coding a handful of common sequences, and fail to correctly ignore sequences they don't implement. And worse, many go on to implement sequences that cannot be correctly handled.

One simple example that usually fails: \e!!F. More nasty, however, are the ones that ignore intermediaries and execute some unrelated command instead.

I can't be bothered to pick apart specific terminals anymore. Most don't even know what an IR is.

The problem with XCB is that it's designed to be efficient, not easy. If you're avoiding toolkits for some reason, "so what if I block the world" may be a reasonable tradeoff.

1, Don't target X11 specifically these days. Yes a lot of people still use it or at least support it in a backward-compatible manner, but Wayland is only increasing.

2, Don't fear the use of libraries. SDL and GTK, being C-based, should both be feasible from assembly; at most you might want to build a C program that dumps constants (if -dM doesn't suffice) and struct offsets (if you don't want to hard-code them).

Even logging can sometimes be enough to hide the heisgenbug.

Logging to a file descriptor can sometimes be avoided by logging to memory (which for crash-safety includes the possibility of an mmap'ed file, since the kernel will just take care of them as long as the whole system doesn't go down). But logging from every thread to a single section of memory can also be problematic (even without mutexes, atomics can be expensive and certainly have side-effects) - sometimes you need a separate per-thread log, and combine in the log-reader tool.

I don't remember the last time I used ctrl-C. It's always select or "+y.

True, but successfully doing dynamically-linked old-disto-test-environment deployments gets rid of the real reason people use static linking.

DNS-over-TCP (which is required by the standard for all replies over 512 bytes) was unsupported prior to MUSL 1.2.4, released in May 2023. Work had begun in 2022 so I guess it wasn't EWONTFIX at that point.

Here's a link showing the MUSL author leaning toward still rejecting the standard-mandated feature as recently as 2020: https://www.openwall.com/lists/musl/2020/04/17/7 ("not to do fallback")

Complaints that the differences are just about "bug-for-bug compatibility" are highly misguided when it's useful features, let alone standard-mandated ones (e.g. the whole complex library is still missing!)

I haven't managed to break into the JS-adjacent ecosystem, but tooling around Typescript is definitely a major part of the problem:

• following a basic tutorial somehow ended up spending multiple seconds just to transpile and run "Hello, World!".
• there are at least 3 different ways of specifying the files and settings you want to use, and some of them will cause others to be ignored entirely, even though it looks like they should be used.
• embracing duck typing means many common type errors simply cannot be caught. Also that means dynamic type checks are impossible, even though JS itself supports them (admittedly with oddities, e.g. with string vs String).
• there are at least 3 incompatible ways to define and use a "module", and it's not clear what's actually useful or intended to be used, or what the outputs are supposed to be for different environments.

At this point I'm seriously considering writing my own sanelanguage-to-JS transpiler or using some other one (maybe Haxe? but I'm not sure its object model allows full performance tweaking), because I've written literally dozens of other languages without this kind of pain.

WASM has its own problems (we shouldn't be quick to call asm.js obsolete ... also, C's object model is not what people think it is) but that's another story.

At this point, I'd be happy with some basic code reuse. Have a "generalized fibonacci" module taking 3 inputs, and call it 3 ways: from a web browser on the client side, as a web browser request to server (which is running nodejs), or as a nodejs command-line program. Transpiling one of the callers should not force the others to be transpiled, but if multiple of the callers need to be transpiled at once, it should not typecheck the library internals multiple times. I should also be able to choose whether to produce a "dynamic" library (which can be recompiled later without recompiling the dependencies) or a "static" one (only output a single merged file), and whether to minify.

I'm not sure the TS ecosystem is competent enough to deal with this.

The problem is that the application developer usually thinks they know everything about what they want from their dependencies, but they actually don't.

The problem is that GLIBC is the only serious attempt at a libc on Linux. The only competitor that is even trying is MUSL, and until early $CURRENTYEAR it still had worldbreaking standard-violating bugs marked WONTFIX. While I can no longer name similar catastrophes, that history gives me little confidence.

There are some lovely technical things in MUSL, but a GLIBC alternative it really is not.