kbin.social has been totally down for a while. I don’t think your posts are actually federating when you post into a kbin.social magazine right now; the votes you are getting are probably from other lemmy.world users only.

I was curious, so I did some searches on this topic for you and found these pages:

• https://news.ycombinator.com/item?id=18155134
• http://www.catb.org/esr/faqs/practical-python-porting/

The second link in particular notes:

The reason that things are much easier with all ASCII data is that practically every Unicode encoding in existence maps bytes 0x00…0x7f to the corresponding code points, so byte strings and Unicode strings that contain the same all-ASCII data are basically equivalent, even semantically. What usually trips people up with non-ASCII data is that the semantic meaning of bytes in the range 0x80…0xff changes from one encoding to another.

But, thinking like a systems programmer again, for many purposes the semantic meaning of bytes 0x80…0xff doesn’t matter. All that matters is that those bytes are preserved unchanged by whatever operations are done. Typical operations like tokenizing strings, looking for markers indicating particular types of data, etc. only need to care about the meaning of bytes in the range 0x00…0x7f; bytes in the range 0x80…0xff are just along for the ride.

So the trick for beating Python 3 strings into submission is to put in encoding and decoding calls where you need to, choosing a single-byte encoding that doesn’t mutate 0x80…0xff. There are many of these; most of the Latin-{1…6} sequence (aka ISO-8859-1…10) is has this property. What you do not want to do is pick utf-8 or any of the multibyte Asian encodings. Latin-1 will do fine; in fact it has an advantage over the others in memory consumption, which we’ll describe below.

Whether depending on this is actually correct or not is beyond me, but it seems like people have actually been using that pass-through behavior in practice and put it into things like Python2 -> 3 migration guides.

The first link suggests that the seemingly undefined ranges are valid as C0 and C1 control codes which may be why it doesn’t throw errors.

Now imagine if to buy a car you had to tolerate cameras and other forms of tracking your telemetry just to get to work and feed yourself.

Sorry to be the bearer of depressing news, but that’s basically already happening in new cars.

https://foundation.mozilla.org/en/blog/privacy-nightmare-on-wheels-every-car-brand-reviewed-by-mozilla-including-ford-volkswagen-and-toyota-flunks-privacy-test/

https://jacobin.com/2024/03/car-spying-insurance-surveillance-data/

I wonder what cuil things it will say if you start asking questions about hamburgers instead…

I ran into an example of the thumbnail issue again today – this time on a post from kbin: https://old.reddthat.com/post/19193476

The thumbnail looks like this in the HTML:

<div class="thumb">
  <a class="url"
     href="https://media.kbin.social/media/60/a4/60a45b8ff88b1b2e3a0f77b701feb323c5bbfb7ceeb75154ea7df5d6eea15ef8.jpg"
     >
    <div  style="background-image: url(https://media.kbin.social/media/60/a4/60a45b8ff88b1b2e3a0f77b701feb323c5bbfb7ceeb75154ea7df5d6eea15ef8.jpg?format=jpg&amp;thumbnail=96)"></div>
  </a>
</div>

Note that it’s making a request to kbin.social with ?format=jpg&thumbnail=96 parameters in the CSS – which results in the full image being loaded since kbin doesn’t run pictrs.

The versions in use on reddthat (according to the settings page) are:

lemmy: 0.19.4-beta.7

mlmym: 0.0.44

Mrs Bighead?!

…

*hangs up the phone*

DOSBox runs on both Linux and Windows (and probably Mac too?); I was suggesting it since you might be able to replace the dying DOS computers with a modern system and just launch the legacy system as an application under it. (You might be able to do the same with a VM as well, but DOSBox came to mind first and may be easier to setup and distribute.)

Just a thought. If it’s not useful, feel free to disregard.

Can you run the DOS software under DOSBox?

Personally, I prefer it when people do one of the following:

• upload the file to catbox.moe and link it here (for clips up to 200MB)
• upload the file directly to their lemmy instance (if their instance allows it)
• self-host it on their own web server (with no bullshit crappy JS interface, please – just give me the file; I’ll play it with VLC if it doesn’t work in my browser)

PeerTube is also a reasonable choice – although I don’t like its UI very much.

I’m having trouble finding an example of the thumbnail issue again right now but I was seeing the pictrs conversion parameters passed to URLs from catbox.moe, i.postimg.cc, and other sources in the CSS for the thumbnail when I reported the issue to Tiff ~3 weeks ago. It’s possible that it got fixed/suppressed by another change since then though. (0.0.44 was deployed a few hours ago and I think there may have also been a beta patch bump for the lemmy backend at some point since I reported the issue originally in our local support community.)

I’ll let you know if I see it pop up again.

For the text handling issue, I was seeing text like “<thread title> by <username> in <community>” (i.e. "<thread title> by <username> in <community>" if it still happens) getting misinterpreted as raw HTML instead of being escaped. (i.e. <!-- raw HTML omitted --> was showing up in the HTML output for the page.)

You may recognize that text as the pattern for a recently fixed bug in the user profiles; I found the text handling issue while trying to explain the other issue to Tiff a few weeks ago.

Will edit this comment immediately after posting to let you know if I still see the text issue.

EDIT: I still see the text issue show up in this comment. https://old.reddthat.com/comment/10370610

Thanks! I’ll go ask Tiff about getting reddthat updated later.

BTW, is there a community for discussion of mlmym itself somewhere on lemmy? I can’t participate on GitHub, but those aren’t the only issues I’ve found. (e.g. there’s also ?format=jpg&thumbnail=96 on non-pictrs links and a text handling issue with angle brackets…)

That’s great to hear!

You can open any profile with multiple pages worth of posts or comments on old.reddthat.com and it’s jumbled. Even my own profile is jumbled: https://old.reddthat.com/u/e0qdk

The first page is mostly comments I made two weeks ago plus a thread from today and some very old threads. The second page has comments I made earlier today and during the past week. The third page starts with my most recent comment and then has a bunch of older comments.

The exact order might change after posting this, but my own recent comments mostly being on page two has been pretty consistent for a while.

If I look at a very active user’s profile (like MentalEdge’s), I see threads from today show up on page three(!) while there are threads from a week or more ago on pages one and two.

I’m not sure what’s going on exactly, but it basically makes user profiles pretty useless right now through mlmym.

Edit: I can’t even find this comment in my profile, but my other reply (regarding the envelope being fixed in 0.0.43) shows up on page 3.

I haven’t had much issue with lag, generally, but I don’t get notifications any more – which is probably the most pressing issue. (I have to remember to manually check once in a while after I post since the envelope doesn’t light up.) That might be an issue with reddthat being on a recent beta version of lemmy – I don’t know.

We do have lemmy-ui-next over here too. Thanks for reminding me about that. I’ve been meaning to poke at it a bit.

Right now I’m mostly using mlmym (the “old” interface on most instances that support it) because it doesn’t require JS for basic viewing.

It’s kind of buggy though, unfortunately – things like user history show up as a complete jumble, for example. :(

One of these days, I’ll probably get fed up enough to go write my own interface and set things up exactly how I want them to work… but I’ve got too many projects already so I’m just living with it for now.

I wonder if this will actually cause an increase in the number of security vulnerabilities and breaches as there’s now a fairly obvious way for employees to penalize their bosses financially for being assholes…

I don’t know about KDE in particular, but I’ve had problems with USB mice waking various Ubuntu systems when they’re not directly connected (i.e. there’s a hub or KVM in between it and the computer). The workaround I used for that was to remove the mouse input (e.g. by carefully pressing a physical button on the KVM) – which was good enough for me – but I think there is a programmatic way to block particular classes of input from waking the system if some device is waking your system inappropriately.

Doing a quick search turned up this: https://askubuntu.com/questions/252743/how-do-i-prevent-mouse-movement-from-waking-up-a-suspended-computer – I can’t vouch for any of the specific techniques there though.

Worth noting that while I had a problem with the mouse specifically, other hardware could be causing your system to wake up.