Nah, you can have a license that says you get a cupcake and another license that says you need to give up your first-born.
And then you can mush those licenses to say that you need to give up your first-born, but you get a cupcake in return.

Unless the specific license terms contradict, this is totally possible.

AGPL is specifically for web services. For example, if Nextcloud were provided under the GPL, Amazon or the like could serve a modified version of Nextcloud without having to hand out their modifications. As far as the GPL is concerned, Amazon is the user and the software just happens to accept requests from the network.
With AGPL, those who use the software over the network are also deemed users and therefore have the right to access the source code.

I imagine, the scenario would be that the cloud service links against a library under the supposed new license.
And then, even if you’re just using the cloud service over the network, you can demand changes to the source code of that library to be open-sourced.

I imagine, it’s just too much of a niche and practically not enforceable anyways.

You would need to somehow know that a web service is a using a modified version of your library, then you’d be able to demand those library changes to be open-sourced.

And well, just in general, covering all kinds of niche use-cases isn’t terribly healthy for open-source licenses, because each modification is something that can be challenged in court and which might be incompatible with other licenses.
Ultimately, a library under such a specialty license would probably not see much use either. You could only really depend on it in AGPL applications. And at some point, you do have to ask yourself, if it’s even useful to develop your library then.

Had a refinement yesterday, where we decided that we should add all tickets of an epic individually into the milestone (except for two).
And for whatever reason, our project manager had decided to use the in-browser split view and was struggling against that, but also just was about to do it in some cumbersome way. I think, he wanted to manually compare the list of issues in the epic vs. the milestone.

Either way, I could tell that he’d need 10+ seconds to even get started. And telling him how to do it would probably take equally long. So, I just open each issue of the epic in a new tab and check on each tab that the issue is in the milestone or add it, then close the tab. And yep, I was long done when he was still trying to find the issue list for the milestone.

That was certainly one of those moments. 🫠

He isn’t entirely familiar with that issue tracking UI, so it’s fine, and of course, it is my job to be good with computers and all that, but still felt wild that he could’ve easily needed ten times as long to do the same thing.

Last week, some LLM bot commented under one of our issues and it became apparent pretty quickly, that it is a bot. So, I went to report it (incredibly the report menu did say they want reports for bots).

I filled out the reporting form probably five times in total, trying at different times of the day. Every time, I got an error 500 (Internal Server Error) as response.
Later, I checked my mails, and saw that actually two of my reports did go through, meaning I created two tickets on their side.

What those mails also said: They’re very sorry, if it takes longer, since they’re currently experiencing a higher number of reports.

Gee, I wonder why.

It’s a reference to another Microsoft classic: https://www.windowscentral.com/software-apps/microsoft-caught-plagiarizing-graphics-with-ai-slop-microsoft-continvoucly-morged-my-diagram-there-for-sure

Hmm, for whatever reason, I’m on 2.31.4, so that might be the difference.

That version was tagged two weeks ago, because they apparently still release patch versions for rather old minor versions of nix. So, apparently I am getting updates, but I’m on some older release channel or something. No idea why.

I have to head to work now, so will have to debug in the evening or the weekend. Thanks for the clue, though.

I don’t think, we are doing different things. I create a new file, put {} inside, then add it into the imports = [...];. It gives me that error.
Then I git add ., run again and don’t get the error anymore.

Is the error you pasted now from some manual assertion you did?

Hmm, that sounds exactly like my setup. Weird.

I did have the file created, with {} inside (empty Nix expression). If I git add it, it works as well:

And yeah, I understand that it’s supposed to be a stacktrace, but other error messages look similarly horrendous and I can often only try to guess what’s wrong by reading the stacktrace top-to-bottom, so I’ve somewhat gotten used to doing that.

But good to know that these terrible error messages might be a problem with my system. Thanks!

Hmm, that’s interesting. For me, it looks like this:

I actually thought, it said somewhere in there, that the file isn’t staged, but apparently not even that (anymore?).

You don’t happen to be using Lix or something, do you? I’ve heard that it’s supposed to have better error messages, but I was never sure how much better it might be…

Edit: Perhaps I should add that those code locations it shows, are not from my code. Only the modules/terminal/new_file.nix in the second-last line is relevant.

I thought, you posted about the warning, because that’s actually easier to see than the error. Because yeah, it does say what you posted, but it’s in the middle of like 30 lines of other stuff. When I forget to stage a new file, it almost always takes me 5+ seconds to spot what the problem is. 🥲

Unfortunately, that shows up even when you’ve just modified an existing file, which is not a problem for it.

And which also happens to be the state my repo is in basically all the time, because I’ll change some setting, then see if it works like I want it to before making a commit…

Yeah, I was gonna say, that might be the root cause.

In the vast majority of cases, you want Box<dyn Error + Send + Sync>, but folks tend to leave out the Send + Sync, because it looks like additional complexity to them, and because it doesn’t cause problems when they’re not doing async/await.
It’s better to define a type alias, if you don’t want that long type name everywhere.

Nice, that’s like the meme:

Look at how quickly AI put up a webpage for me: http://127.0.0.1/index.html

until some random AI agent

Wait, do they now have spam bots going around on random PRs to post advertisements?

I’m not much of a fan of Debian, but in your position would still recommend it. You’ll have enough to learn about from just using it as a server. You can learn about potential advantages of other distros later…

Gab vorher kein Firefox Maskottchen. Es gibt mehr oder weniger noch ein Mozilla Maskottchen mit dem ursprünglichen Dino-Logo, falls du das im Kopf hast:

It’s right-wing trolling that it’s specifically non-binary. It’s just iconography they use throughout Firefox, when displaying error messages or the like.

Okay, but just to be clear, the problem is not that it can’t do a timer. The problem is that it claims to be able to and even produces a result which looks plausible. It means, you cannot trust it to do anything that you can’t easily verify. If they could fix that overconfidence in a year, it would be much better.