@Sumomo@sumomo.neocities.org You should report the bug and hopefully get it fixed 🤞

@lyse@lyse.isobeef.org But I wannaaaaaaaaaaaaaaaaaaa. 😭

Let’s see how this goes: https://bugzilla.mozilla.org/show_bug.cgi?id=2065883

@brytboi@twtpub.com I hope you’re seeing my replies because you absolutely can scroll up in the app. Let me know if you’ve run into a bug though and report it to me so I can fix it immediately!

@david@daiwei.me Are you able to describe this bug as an Issue in the Gitea Issue Tracker such that it can be reproduced and fixed? 🙏

Data protection, @dce@hashnix.club! :-D

This revealed another bug in my client that I still need to fix at some point. Inserting an empty set of messages failed with an SQL logic error. Whoops. I didn’t think about that corner case.

@movq@www.uninformativ.de Finally, your software is just perfect and finished by now, no need to report non-existing bugs or send in code changes. :-) How many tickets and merge requests did you get before moving to your own server?

I have to admit that I use git format-patch so rarely, I always have to pull it up from my shell history. Haven’t used git send-email even once. I definitely have to look into that soon. Wanted to do that for several years. I typically upload the patch to my server and send a link via IRC.

Maybe I was just very unlucky, but my experience is that you can perfectly ignore people and their work who only do it for the “fame”. It’s almost always been from inferior quality to say the least.

@prologic@twtxt.net Hmmm, I have no idea how to solve that problem. 😅 Some jenny stuff aside, I received zero bug reports or code contributions since leaving GitHub in 2018.

I thought I had made that super easy, because you can just send me an email – no sign-up process, nothing. But that’s way too old-school, people don’t know how to use git format-patch (let alone git send-email) and they also don’t understand that they can just send me a link to their forked Git repo (which can be hosted anywhere). Git is super flexible and powerful, but those features are hardly ever used.

Maybe people even need some kind “reward”, or “fame”. Like those “achievements” that you can unlock on GitHub. (Something to put in their CV … ?)

Okay, so, my website also includes my code / git repos, and those are made browsable by stagit. What I don’t like about this (these days) is that this includes all the diffs of my commits. In other words: All my code.

This makes it super easy for malicious crawlers to slurp up valuable data. I don’t like that.

I’m thinking about switching to this instead:

https://movq.de/git2/

It still shows project infos and there are Atom feeds, but to get the code, you have to actually clone the repos.

🤔

(If you spot any bugs, let me know.)

@lyse@lyse.isobeef.org Ahh ok! I was just wondering and curious whether it was a bug that I’ve caused anywhere along the way 🧐

@prologic@twtxt.net A screenshot won’t help in this case, as you don’t see anything. :-D It starts off just fine with a conversation tree like that:

Unknown conversation root
└╴Read reply
  └╴Read subreply

Everything works. After reloading the feeds, a new message becomes part of the conversation, so the conversation e.g. looks:

Unknown conversation root
└╴Read reply
  ├╴Read subreply
  └╴Unread subreply

However, the bug is that the whole conversation is not shown at all. None of the three (or four with the root) messages appear in the message tree view. My recursive SQL determining the messages to display is clearly broken.

Ahh yes! Please do upgrade your twtd instance. Few things changed, many bugs fixed there too.

I should really fix this damn bug where new replies to read replies to unknown conversation roots are not showing up in tt. I run into this bug almost daily for weeks now.

@dce@hashnix.club Fixed! 🥳 That 403 was our bug — connect was checking your token via /api/v1/user (needs read:user), but your token’s scoped to just the repo so it can’t. Now it validates against the repo itself instead 👍

@david@daiwei.me I think it might be a bug i just fixed 🤞

@david@daiwei.me Found it. Some bugs in the “claim limiter”. Fixing…

based on this, it’s entirely possible that there may still be a subtle bug somewhere with the app

@movq@www.uninformativ.de please don’t waste your time to bugging this. I’ll figure out what’s going on with these new clients.🙏

Hurray, I can now press gg instead of g to go to the top in tt. Much better! :-) Other multi-key combinations are also easily possible now.

I should probably write a real article about this at some point, but here we go. The only downside with my new key binding system is that it breaks tview’s established pattern. You’ve got an InputHandler(), that is implemented using WrapInputHandler(…). It typically then directly implements the switching logic depending on the key press. Something like this:

func (w *Widget) InputHandler() func(event *tcell.EventKey, setFocus func(p tview.Primitive)) {
    // WrapInputHandler allows for intercepting key events with SetInputCapture(…)
    // from the outside for customization. This handles the default key bindings.
    return t.WrapInputHandler(func(event *tcell.EventKey, setFocus func(p tview.Primitive)) {
        switch event.Key() {
        case tcell.KeyRune:
            if event.Modifiers() == tcell.ModNone {
                switch event.Rune() {
                case 'k':
                    w.scrollUp()
                    return // we already handled the event, stop processing

                case 'j':
                    w.scrollDown()
                    return
                }
            }
        }

        // We didn't handle the key event. Maybe the parent
        // widget knows what to do with it.
        if handler := w.parent.InputHandler(); handler != nil {
            handler(event, setFocus)
        }
    })
}

From the outside, you can intercept and either stop or continue the widget’s original key handling with a potentially rewritten key event using SetInputCapture(…):

w := NewWidget()
// customized or additional key bindings
w.SetInputCapture(func(event *tcell.EventKey) *tcell.EventKey {
    switch event.Key() {
    case tcell.KeyUp:
        // Rewrite the event, so the "cursor up" key is an alias
        // for the vim key binding "k", that is handled by the
        // wrapped input handler above. (I know, I know, this is a
        // completely unrealistic example, why would anyone use
        // cursor keys when there are vim key bindings available?!)
        return tcell.NewEventKey(tcell.KeyRune, 'k', tcell.ModNone)

    case tcell.KeyRune:
        if event.Modifiers() == tcell.ModNone {
            switch event.Rune() {
                case 'q':
                    app.Stop()
                    // we already handled the event, do not pass it
                    // to the wrapped input handler above
                    return nil

                case 'r':
                    toggleMessageReadStatus()
                    return nil
            }
        }
    }

    // we didn't handle the event, pass it to the wrapped
    // input handler above
    return event
}

Since they all expect a single key, I’ve noticed that using multiple dedicated KeyBindings of mine on these different levels kinda breaks multi-key handling with common prefixes. The outer-most KeyBinding captures the prefix, but it can’t transfer it to the inner one if not handled by the outer one. At least not without some more (potentially ugly) changes. So, I now have to work with just a single KeyBindings object for the entire widget chain (if it consists of multiple other widgets or the regular input handler and input capture are in the game). The outside needs to register all its key bind customizations or extensions at the same level that the original widget handles its default ones. Doable by exposing the widget’s KeyBindings instance, but not pretty. You always have to keep this in mind.

With the KeyBindings, it will look like that:

type Widget struct {
    parent tview.Primitive

    // make it available to children or the outside either by
    // direct field access or by providing a getter method
    KeyBindings *bind.KeyBindings
}

func NewWidget() *Widget {
    w := &Widget{KeyBindings: &bind.KeyBindings{}}
    w.KeyBindings. // default key bindings
        Bind0(bind.KeySequence('k', w.scrollUp).
        Bind0(bind.KeySequence('j', w.scrollDown)
    return w
}

func (w *Widget) InputHandler() InputHandler() func(event *tcell.EventKey, setFocus func(p tview.Primitive)) {
    return t.WrapInputHandler(func(event *tcell.EventKey, setFocus func(p tview.Primitive)) {
        // also note the missing support for focus transfer at the moment
        event = w.KeyBindings.Capture(event)
        if event == nil {
            return
        }

        if handler := w.parent.InputHandler(); handler != nil {
            handler(event, setFocus)
        }
    }
}

And then from the outside, or in a child widget:

w := NewWidget()
w.KeyBindings. // additional or customized key bindings
    Bind1(bind.KeySequence(tcell.KeyUp), func(*tcell.EventKey) *tcell.EventKey {
        return tcell.NewEventKey(tcell.KeyRune, 'k', tcell.ModNone)
    }).
    Bind0(bind.KeySequence('q'), app.Stop).
    Bind0(bind.KeySequence('r'), toggleMessageReadStatus)

When directly working with tview primitives that are not part of custom widget implementations, the following works well so far:

textView := tview.NewTextView().
    SetWordWrap(true).
    SetText("…")
    SetScrollable(true)
textView.SetInputCapture((&bind.KeyBindings{}).
    Bind0(bind.KeySequence('q'), app.Stop).
    Bind1(bind.KeySequence('g', 'g'), func(*tcell.EventKey) *tcell.EventKey {
        return tcell.NewEventKey(tcell.KeyHome, 0, tcell.ModNone)
    }).
    Capture)

I need to sleep on this some more.

Also, writing very long messages like this one is really not all that fun in tt’s editor. I should absolutely provide a way to shell out to vim.

(Took me about one and a half hours to compose, holy crap. But not only because of not using vim. Although, that might have saved me a quarter hour or so for sure. Proof-reading this message also uncovered quite a few bugs in my real documentation. So, that’s a big win!) Good night!

I believe we’ve nailed all the bugs down🤣 Though i am sick at the moment so i’m not at my best 😢

@balloon-fu-sen@tw.fus.f5.si Ahhh! That’s a bug! Lemme fix that!

🥳 Finally! After nearly 4 years, yarnd v0.16.0 “Silver Sojourner” is out! 🚀 Twt Hash v2, SQLite FTS5 search, HTMX-powered UI, first-time setup wizard and literally hundreds of bug fixes 🐛

Release notes: https://git.mills.io/yarnsocial/yarn/releases/tag/0.16.0

Upgrading is fully automatic — the Twt Hash v2 migration re-fetches all feeds on first start, so expect the first cycle to be a bit heavier. Images on Docker Hub as prologic/yarnd:0.16.0 👌

cc @kat@yarn.girlonthemoon.xyz @abucci@anthony.buc.ci @shinyoukai@yume.laidback.moe @eldersnake@we.loveprivacy.club 🙏

@bender@twtxt.net No idea. I can only tell you that the correct hash would have been rwzz277nkyju for this line:

[2026-07-11 14:47:17+00:00] [(#5bpwpdcjnhcz) <a href="https://txt.sour.is/external?uri=https://daiwei.me/twtxt.txt">@david<em>@daiwei.me</em></a> (This thread is broken again on my end. Another bug or fix not released yet? 😅)]

@david@daiwei.me (This thread is broken again on my end. Another bug or fix not released yet? 😅)

I will aim to have most issues bugs and user experience problems, identified and fixed by this weekend!

And if we can compile a list and file issues for feeds, twtxt.app and anything else as issues for when i get back 🙏 feature requests, bug reports. etc 🤞

FYi 👋 I’m aware of an optimist precomputed hashing bug on the new twtxt.app 🤯 Trying to work with @bender@twtxt.net remotely on my vacation yo fix it 🤣

@movq@www.uninformativ.de I figure there is a bug somewhere, but where?

What’s the bug?

@movq@www.uninformativ.de we found a bug. @prologic@twtxt.net loves it!

I think I fixed this bug!

@lyse@lyse.isobeef.org Found it and fixed it! 🎉 The crawler’s discovery spider was fetching every feed a second time, without any conditional headers (plus a couple of other politeness bugs: redirected feed URLs never stored their cache validators, and there was no floor between re-fetches). Now every feed is fetched at most once per crawl, always with If-Modified-Since / If-None-Match, and never more than once per 15m no matter what. Just deployed — please keep an eye on your access logs and let me know if you still see anything impolite from the crawler 🙏

@lyse@lyse.isobeef.org Thanks! I’ll look into that! Could be a bug in the crawler.

Hey folks 👋 Today I announce the re-release of the Twtxt Search Engine now live and running and actively re-crawling and re-indexing. 🎉 Please report bugs or any useability issues to me! 🙏 #Twtxt #Search

My mate and I hiked up the backyard mountain. We got 25°C and quite some wind, so it was actually not too terrible. The wind could have blown harder or the temps a little lower, but oh well.

I saw the squirrel’s bushy tail stick up on the forest floor in the sunlight and immediately thought of this cute little feller. Since it didn’t move at all, even when we came closer, I got irritated and reconsidered that it might actually be some kind of dried up farn. But then we also were able to see its body. Unfortunately, the squirrel ran up the tree too quickly, so all the shots are kinda crap.

At one flower spot, there were sooo many butterflies, wasps, flies, bugs and other insects. The botanic was completely crowded.

The workers were transferring logs from one log truck to the other in a parking lot. I’ve never seen this happening before. When we passed the same place on the way home, they had moved logs into a sea container. That was surprising. This semi wasn’t there on the way there. One log was probably too long and sticking out the container, so they probably had to wait for somebody to return with a chainsaw. Crazy that they’re shipping logs from here probably overseas. Why else would they put them in a sea container?

After our first break, a blackbird was really posing for us with his worm in the dark shade.

Today was my first time I ever saw a hummingbird hawk-moth (Taubenschwänzchen) for real. My mate photographed them many, many times before, but I never came across one myself. So, that was really special.

The forest service installed an outdoor table with two benches next to the timber lion, that was cool to see. We sat down for a few minutes and enjoyed both the view into the Fils valley and ant on the tabletop, but the sun was beating down too heavily on us, so we had to move on.

All in all, it was a very nice few hours long hike. Enjoy! https://lyse.isobeef.org/waldspaziergang-2026-07-03/

@lyse@lyse.isobeef.org oh yes! And, when I mow the lawn (which reminds me I need to mow the front soonish), you can add dust, bugs, and grass blades to the equation. Just “lovely”. 😂

@lyse@lyse.isobeef.org If I were to guess: They might have done so to avoid bug reports from users with heavily outdated versions. 🫤

@lyse@lyse.isobeef.org Interesting approach. 🤔

The master branch should never be in a broken state (apart from bugs I don’t know about). Any intermediate state during the development of a larger feature will happen in a different branch.

I mean, yeah, but … I don’t know, I like having “traditional releases” as a second safety net when I write programs. I like to let things mature for a while and then I cut a new release. So it’s, like, “we have a bunch of new features and fixes here, and to the best of my knowledge this works fine now”. But maybe I’m just paranoid. 🤔

The FCC Wants to Kill Burner Phones
Plus: AI bug hunting fuels Microsoft’s biggest-ever Patch Tuesday, ShinyHunters ransomware gang exploits an Oracle zero-day, and more. ⌘ Read more

@movq@www.uninformativ.de Oh yeah, way better! :-) I didn’t spot the bug, though.

I think I could work with the feature set. I typically don’t need a lot. Until I do. :-D The message tree in tt is an example of that. But tt is also special that it needs something like this in the first place. It’s unusual.

(And of course there’s a bug because I’m an idiot. 🤪)

CISA Tells US Agencies to Fix Security Bugs in as Little as 3 Days Thanks to AI Threats
“Defenders cannot afford to take weeks to patch,” one Cybersecurity and Infrastructure Security Agency official warned on Wednesday. ⌘ Read more

When a user reports a bug in a feature I didn’t even know we had ⌘ Read more

@itsericwoodward@itsericwoodward.com Turns out, this is a bug in my config to cache synchronization. Nickname changes in the configuration file are just not synced to the cache at startup if the feed URL already exists in the cache. I must have fixed this typo in my config ages ago, because I don’t even recall having that spelling mistake to begin with. Yet, the cache was happily showing the erroneous nickname. Composing a reply automatically adds the mentions from the conversation participants. Everything originates from the cache, so, I successfully poissoned my replies.

@movq@www.uninformativ.de Honestly I think you build the team before you need the PRs 🤔 Start with relationships — people who’ve been using your software, filing good bug reports, asking smart questions. Those are your future maintainers. The PR comes later as a formality, not a tryout 😅

Ask HN: Why is the HN crowd so anti-AI?
Genuine question.

Over the past six months, there hasn’t been a single day where I’ve checked the HN Best RSS feed without seeing a post about how AI “writes bad code,” “introduces bugs,” “creates technical debt,” or something along those lines.

I’ll probably make a lot of enemies by saying this, but do people realize that code is just a means to an end?

Users don’t care whether the code was written by AI or by hand, or which framework you used. They care that the product works.
… ⌘ Read more

Not to Alarm Anyone, but Flesh-Eating Screwworms Have Entered the US
The USDA this week confirmed the first known infection of the carnivorous fly larva, which feast on the flesh of living mammals, after the United States eradicated the nightmare bugs in the 1960s. ⌘ Read more

When QA reproduces in two clicks the bug I swore was impossible ⌘ Read more

Yay finally fixed some of those annoying “Mark as Read” behaviours/bugs 🐞

I wasted my entire weekend on the writeup. If you have way too much time to spare and also are interested in a bug analysis of a software that you don’t even use, I have you covered: https://lyse.isobeef.org/newsboat-time-parsing-bug-analysis/

Oh boy, it was bloody humid this morning. Just around 20°C when we left, but climbing rapidly. The flow of air when walking was okay, but as soon as we stopped, streams of sweat were pouring down on us. Luckily, it was cloudy, but the lack of wind was bad. Now, the sun is out, 29°C will be reached in an hour and I’m glad that the house is still cool. It will be a different story in a few weeks or months. Not looking forward to that at ll.

On the bright side, we saw the first tadpoles of the year and an also first, but sadly dead slow worm that probably some bird dropped on a bench next to the fountain. The fly was stuck to its feast and also cactus. The municipality fixed the railing nicely and we came across a giant patch of great looking fire bugs on the summit.

All in all, a successful stroll through the woods but for the humid heat.

https://lyse.isobeef.org/waldspaziergang-2026-05-30/

@prologic@twtxt.net Ahh, I see. Okay, I’m with you there. On this high level, I can understand how the thing works.

Maybe my wording isn’t good. 🤔 Let’s take a real life example from what we do at work.

There’s this AI chatbot. It gets support requests from users, so the user says something like “I need access to a particular system”. This triggers the bot to “run” the instructions stored in a large Markdown file, like “check if the user is authorized to do this, then issue the following API requests”, and so on. This is essentially like running a little script, except it’s written in natural language (German) and there’s no “script interpreter” but just the AI.

Now, suppose that the AI doesn’t quite do what was intended. There’s some subtle bug. How do you debug this? How do you find out how the AI came to the “conclusion” to run step A instead of step B? And how do you find out how exactly you have to change your prompt so this doesn’t happen again next time?

If this was an actual script/program instead of AI, you could repeat the request and attach a debugger or throw in some printf() or whatever. How do you do that kind of thing with AI? How do you pinpoint exactly what the problem was?

(Or is this just a stupid idea? Do we have to give up that way of thinking when using AI? Is the era of debuggability over?)

Aha, yesterday’s newly added support for LC_TIME to render localized timestamps also broke the feed parsing with my LANG=de_DE.UTF-8 and LC_CTYPE=de_DE.UTF-8 environment. :-)

Atom feeds make use of RFC 3339 timestamps. They are first converted into RFC 882 timestamp representation, which is the one that RSS feeds use. However, this conversion now results in localized RFC 882 timestamps, which cannot be parsed into Unix timestamp numbers via curl_getdate(…). I bet that it doesn’t know about the localization at all and expects English month and weekday names. Looking at its docs, I reckon that function was selected because of its myriad of supported timestamp formats: https://curl.se/libcurl/c/curl_getdate.html RFC 3339 is not included, though, hence the transformation up front.

The intermediate Item objects in the parser domain use std::string for the timestamp representation. This isn’t all that silly, because Newsboat supports all sorts of different feed formats with different timestamp formats. These RFC 883 timestamps are centrally parsed into time_t.

Speaking of time: It’s time to go to bed after this late bug hunting fun. :-)

You didn’t change your Atom feed by any chance yesterday or today, @movq@www.uninformativ.de? Not only do I have a metric shitton of “new” old items in my YouTube feeds, but also a bunch of your old articles are shown as new.

I fear that this is a Newsboat bug. I rebuilt it yesterday from master.

The AI Era Is Creating a Bug Hunting Arms Race
As attackers ramp up their AI exploit development, the search for software vulnerabilities is changing rapidly. ⌘ Read more

When I finish my dev work and testing flags 10 bugs within the hour ⌘ Read more

When I’m the only one who notices we have a huge bug in our demo ⌘ Read more

When I ask Claude Code to fix a bug but forgot to give it access to my repo ⌘ Read more

Best Bug Spray (2026), Tested and Reviewed
Our writer tried some of the most popular insect repellents during runs, hikes, and evening walks. These are the ones we recommend. ⌘ Read more

When we’re celebrating the production release and I suddenly remember a forgotten bug ⌘ Read more

When I show the sales rep how to work around a display bug for their demo ⌘ Read more

When I follow the instructions to reproduce an exotic bug ⌘ Read more

Discord Sleuths Gained Unauthorized Access to Anthropic’s Mythos
Plus: Spy firms tap into a global telecom weakness to track targets, 500,000 UK health records go up for sale on Alibaba, Apple patches a revealing notification bug, and more. ⌘ Read more

When I’ve been searching for hours and still don’t understand where the bug is coming from ⌘ Read more

When the lead dev personally goes to the client to fix a major bug ⌘ Read more

When it’s the sixth time ChatGPT fails to fix the same bug ⌘ Read more

When I finally manage to fix a bug after pulling an all-nighter ⌘ Read more

When someone comes looking for me to fix a bug in production ⌘ Read more

When I find a workaround for a bug in production ⌘ Read more

When the client asks if we can remove the bugs but keep the features ⌘ Read more

When the intern fixes in 15 minutes the bug I’ve been stuck on all day ⌘ Read more

When the bug only shows up on the CEO’s browser ⌘ Read more

When QA finds the first bug in my code in 30 seconds ⌘ Read more

When I manage to fix the bug the team had been stuck on for days ⌘ Read more

When the bug disappears when I add a console.log ⌘ Read more

Hmm I think it’s a bug in the Javascript. It’s meant to be ![title](url)

When I realize my hotfix created 3 other bugs ⌘ Read more

When I manage to fix a bug without using Google or an AI ⌘ Read more

When I keep chaining fixes to solve the same bug ⌘ Read more

When the client didn’t notice the major bug during my demo and congratulates me ⌘ Read more

When the bug comes from a non-breaking space in the input string ⌘ Read more

When my pair tells me they found a bug 5 minutes before my demo ⌘ Read more

When the manager tells me the bug is top priority even though it’s been around for 6 months ⌘ Read more

When I manage to reproduce the bug that was reported to me ⌘ Read more

@prologic@twtxt.net (While browsing through that, I noticed that https://mu-lang.dev/ itself doesn’t really mention the source code repo, does it? 🤔 Like, the quickstart guide begins with “Build the host: go build ./cmd/mu”, but where’s the git clone … command? 😅)

I’m not really sure what the goal is. 🤔 Do you want to get pull requests for the docs? Or bug reports for mu itself? 🤔

When I head into the weekend after helping another team fix their bug ⌘ Read more

@bender@twtxt.net That’s the plan! Once I’m happy with this v1 (and we find no other obvious bugs/issues) updating “Changes” with user-facing / human-freidnyl changes is part of the release process!

When a coworker’s code crashes because of a hardcoded “Sure, here’s what should fix the bug:” ⌘ Read more

When I finally understand the bug I was stuck on right after leaving the office ⌘ Read more

When they ask how I fixed the bug but it was actually ChatGPT who did it. ⌘ Read more

When I fix a bug and the behavior gets even weirder ⌘ Read more

Heh I thought I fixed that bug? (is it s abug?!)

@bender@twtxt.net it works fine under jenny. Maybe it is a bug on Yarn?

When I try to explain my bug to the lead dev but he’s already convinced the issue is elsewhere ⌘ Read more

@lyse@lyse.isobeef.org Bah. Yeah, that looks like a bug. Let’s see if this already reported upstream. 🤔

Hurray, I finally fixed another rendering bug in tt that was bugging me for a long time. Previously, when there were empty lines in a markdown multiline code block, the background color of the code block had not been used for the empty lines. So, this then looked as if there were actually several code blocks instead of a single one.

Image

I just fixed another bug in tt where the language hint in multiline markdown code blocks had not been stripped before rendering. It just looked like it was part of the actual code, which was ugly. I now throw it away. Actually, it’s already extracted into the data model for possible future syntax highlighting.

@shinyoukai@neko.laidback.moe Because you might not want to commit all changed files in a single commit. I very often make use of this and create several commits. In fact, I like to git add --patch to interactively select which parts of a file go in the next commit. This happens most likely when refactoring during a feature implementation or bug fix. I couldn’t live without that anymore. :-)

If you have a much more organized way of working where this does not come up, you can just git commit --all to include all changed files in the next commit without git adding them first. But new files still have to be git added manually once.

When the lead dev tells me he fixed a couple of bugs in my project over the holidays ⌘ Read more

@lyse@lyse.isobeef.org I’m toying with the idea of making a widget/window system on top of Python’s ncurses. I’ve never really been happy with the existing ones (like urwid, textual, pytermgui, …). I mean, they’re not horrible, it’s mostly the performance that’s bugging me – I don’t want to wait an entire second for a terminal program to start up.

Not sure if I’ll actually see it through, though. Unicode makes this kind of thing extremely hard. 🫤

Whoo! I fixed one of the hardest bugs in mu (µ) I think I’ve had to figure out. Took me several days in fact to figure it out. The basic problem was, println(1, 2) was bring printed as 1 2 in the bytecode VM and 1 nil when natively compiled to machine code on macOS. In the end it turned out the machine code being generated / emitted meant that the list pointers for the rest... of the variadic arguments was being slot into a register that was being clobbered by the mu_retain and mu_release calls and effectively getting freed up on first use by the RC (reference counting) garbage collector 🤦‍♂️