2025-08-16T19:36:26.928Z
Here's the whole code:
```
private static class SingletonList<E>
extends AbstractList<E>
implements RandomAccess, Serializable {

@java.io.Serial
private static final long serialVersionUID = 3093736618740652951L;

@SuppressWarnings("serial") // Conditionally serializable
private final E element;

SingletonList(E obj) {element = obj;}

public Iterator<E> iterator() {
return singletonIterator(element);
}

public int size() {return 1;}

public boolean contains(Object obj) {return eq(obj, element);}

public E get(int index) {
if (index != 0)
throw new IndexOutOfBoundsException("Index: "+index+", Size: 1");
return element;
}

// Override default methods for Collection
@Override
public void forEach(Consumer<? super E> action) {
action.accept(element);
}
@Override
public boolean removeIf(Predicate<? super E> filter) {
throw new UnsupportedOperationException();
}
@Override
public void replaceAll(UnaryOperator<E> operator) {
throw new UnsupportedOperationException();
}
@Override
public void sort(Comparator<? super E> c) {
}
@Override
public Spliterator<E> spliterator() {
return singletonSpliterator(element);
}
@Override
public int hashCode() {
return 31 + Objects.hashCode(element);
}
}
```
I think my favorite thing here is the hash code just being 31 plus whatever its element's hash code is
sort() being an empty method is pretty good too though
2025-08-18T22:32:52.338Z
catching up on the actual Honey lawsuit:
* PayPal filed a motion to dismiss. I won't cover it because I don't feel like reading it right now
* PayPal also filed a motion to compel arbitration for certain plaintiffs who they contend have PayPal or Venmo accounts. Should be interesting to see how that goes.
* The parties disagree on a discovery aspect: whether the extension source code should be viewed in an online environment (so that it can be analyzed how it interacts with the backend servers) or not (as is standard in source code analysis). My guess is plaintiffs will lose this one, I don't think it's an "undue burden" for them to analyze the source code directly.
2025-08-19T15:26:49.919Z

darn, I wanted to be the bonaafidest
2025-08-20T00:35:41.547Z
apparently Windows has a (registry) setting to disable SSL for time synchronization to improve synchronization accuracy slightly, which is neat
2025-08-22T04:53:05.413Z
Idea: fujoshi/himejoshi but for straight couples
person who gets extremely excited whenever they see a straight couple represented in media
2025-08-25T21:06:51.361Z
none of the above

2025-09-01T19:17:26.784Z
Cursed Java fact: you can have two methods that vary only by return type in a class file, and in fact Java will generate such methods in certain circumstances, even though you can't write such methods
2025-09-02T23:34:00.725Z
okay switching on booleans is kinda cool actually


(examples taken from https://www.youtube.com/watch?v=X0-TGhktFnE)
2025-09-02T23:35:49.223Z
although you could also do

```java
String decideTernary(boolean b) {
return b ? "true" : "false";
}
```

which I think is the pattern I'd prefer as long as the return values are relatively short
2025-09-03T06:13:19.636Z
I was worried that there wasn't any fanfic of the International Prototype Kilogram, but good news https://archiveofourown.org/works/18578899
2025-09-04T04:04:17.774Z
Oh that's nice, journald automatically zstd-compresses the /var/log/journal files
(with filesystem-level compression, so it's transparent to normal applications)
2025-09-04T04:12:41.133Z
actually nevermind that's not filesystem-level compression that's just regular compression
2025-09-05T13:36:43.562Z
Is it just me or is this description nonsense?

2025-09-05T13:38:04.264Z
like after reading it three times I do understand what it's saying. but it's worded in what seems like the most confusing way
like obviously CRN does not stand for column row number, it stands for ceiling row number
whoever made the doc probably was just reading the code and made a wrong assumption about what it meant
and why is one expression split across two rows?
and why aren't the parentheses in the third bullet matched?
2025-09-05T17:54:39.471Z
interesting bit of trivia: the first world record for 5x5 blindfolded is also the slowest time ever recorded (time of 2:34:36, most modern competitions limit you to 1 hour, and the other competitors didn't complete their attempts successfully)
2025-09-11T17:21:43.464Z
We're so back
and I've been neglecting by duties

I'll read (some of) them in a bit, got other stuff to do first
*my duties
2025-09-11T19:02:41.549Z
Quick update on Microsoft: Microsoft makes an interesting argument that the definition of last-click attribution in their complaint is correct, but not the version in their response to the MtD. Personally, I think either wording should favor plaintiffs. There's some other arguments that I think are ultimately immaterial, but I think this one will be argued over a good bit.
2025-09-11T19:04:01.760Z
All the docket entries in Capital One are mostly not actually that interesting. We're in the middle of discovery, there's not any motions, so normally the docket would be pretty quiet. But we've had some status reports, some hearings, a bunch of motions to seal and memorandums in support of such, etc
2025-09-11T19:06:50.056Z
RetailMeNot doesn't have any real action, I'll skip it
2025-09-11T19:21:09.265Z
Wendover case is relatively minor, there's an order on source code discovery finally, it looks reasonable
and I think I'll stop there and do UK Honey another day, I need to get some more stuff done today
2025-09-14T18:20:28.407Z
apropos of nothing it's impressive how much of human knowledge can be compressed into a few tens of gigabytes
2025-09-14T18:21:49.999Z
obviously Wikipedia has done this for a while. textual exports are in the tens of gigabytes compressed, and they have a lot of human knowledge
but I think LLMs are on the whole better at it. like they have a better distribution of information encoded across the domains that are useful to regular people
2025-09-14T18:23:27.171Z
like, if you want to make spaghetti carbonara, I think you're much better off asking Llama that trying to work from the vague, four-sentence description under "preparation" on the arbonara article
2025-09-14T18:26:30.032Z
on the other hand if you want to know where Hubbard, Canada is, I'd go for Wikipedia every time
2025-09-14T18:28:34.875Z
and sure the coverage for LLMs is spotty and compression is lossy and retrieval is expensive. but simply in terms of how much useful information you can get, I think LLMs win handily
2025-09-15T01:38:37.051Z
I love HTML tags in my product names

2025-09-15T16:12:00.326Z
6 distributions in one browser window might be a record for me

2025-09-15T17:07:06.890Z
bad idea: file system that lets you have multiple files with the same name in the same location