kamstrup

joined 2 years ago
[–] kamstrup@programming.dev 10 points 1 week ago (3 children)

As I explained elsewhere there is no official app to change this setting. Users can hack their gsettings.

Support for middle-paste will slowly but surely bitrot and eventually be removed.

[–] kamstrup@programming.dev 15 points 1 week ago (1 children)

No default gnome app will be able to toggle that default. You can hack it in gsettings.

And worse, the fact there is a setting means that only the default will be tested. The feature will slowly but surely bitrot. In a few years we'll see a proposal to remove it entirely. This is how software development works.

[–] kamstrup@programming.dev 28 points 3 weeks ago

Many moons ago I did a project at uni where we implemented elliptic curve cryptography in Java and released it as open source. Unsurprisingly, we had no idea what we were doing. Some years later I get a random mail from someone using it on some embedded system...

I don't want to know, and I fear that ist is paramount that I maintain plausible deniability 😂♥️🙏

[–] kamstrup@programming.dev 7 points 1 month ago

Do it. DO IT

[–] kamstrup@programming.dev 1 points 1 month ago

That depends on what you count as a "test". In some langs/frameworks it is a lot, indeed.

[–] kamstrup@programming.dev 5 points 1 month ago (3 children)

Yeah. Totally agree on this. I spend maybe 3-4h a day reviewing code, and these are my thoughts....

The LLM generated tests I see are generally of very low quality. Perfectly fitting the bill of looking like a test, but not actually being a good test.

They often don't test the precise expected value. As an overly simplistic example: They rarely check 2+2==4. But just assert 2+2>0, or often just that 2+2 doesn't cause an error.

The tests often contain mountains of redundancy. Again, an oversimplified example: They have a test for 2+2, and another for 2+3.

There is never any attempt to make the tests nice to read for humans. It is always just heaps of boilerplate code. No helpers introduced, or affordances to simplify test setup.

Coupling the proclivity for boilerplate together with subtly redundant tests makes for some very poor programming. Worse than I'd expect from a junior, tbh.

And 1500 tests... That is not necessarily a lot! If that is the output of 1 month of pumping out code, I would say bare minimum

[–] kamstrup@programming.dev 14 points 3 months ago

This is just incorrect, sorry to break the news. Most modern electric cars are hardwired to phone home. In most models the surveillance is fused directly into critical components like the fuel pump or the braking system. You cannot just pull out some wires in the dashboard. If you disconnected these things the car is unlikely to work. These details have been covered by people who have worked in the industry

[–] kamstrup@programming.dev 4 points 3 months ago (1 children)

Most devs I know like recursion. Trouble is that many popular languages don't support tail recursion, but throw a stackoverflow error after a few thousand levels. So you have to keep track of max recursion depth manually, and it starts to look like a complicated solution

[–] kamstrup@programming.dev 1 points 3 months ago

Most devs I know like recursion. Trouble is that many popular languages don't support tail recursion, but throw a stackoverflow error after a few thousand levels. So you have to keep track of max recursion depth manually, and it starts to look like a complicated solution

[–] kamstrup@programming.dev 4 points 4 months ago (1 children)

9 times out of 10, this "vibe draft" sends people down a terrible path that they would have never ventured had there been an adult in the room. I swear I review so much code that sets off in the wrong direction because of this, and I am sick of it.

[–] kamstrup@programming.dev 9 points 5 months ago (5 children)

Forgejo supports SSO, and from a quick skim of the diff it looks like they support GitHub and OpenID logins.

 

The Go team is working on a new garbage collector called Green Tea.

 

In the original proof of concept for ranging over functions, iter.Pull was implemented via goroutines and channels, which has a massive overhead.

When I dug in to see what the released code did I was delighted to see that the go devs implemented actual coroutines to power it. Which is one of the only ways to get sensible performance from this.

Will the coro package be exposed as public API in the future? Here's to hoping ♥️

view more: next ›