Rob Napier

@cocoaphony
2.5K Followers
537 Following
4.3K Posts
Swift and Go. Love 'em both. They make me mad in completely different ways. Infosec as required. robnapier.net
Cocoaphony Bloghttps://robnapier.net
GitHubhttps://github.com/rnapier
Twitter (defunct)https://twitter.com/cocoaphony
Is Feedback Assistant having trouble, or is it just me? (I hate messages that give no hint as to what might have happened.)

"Convert this Java to Swift."
chunka chunka chunka. Done.

Very nice. That would have taken me an hour, probably. 1274 lines converted in a few minutes. AI is the best.

And it runs. Excellent.

And the results are good. Awesome.

But… it's slow? My whole point was to implement a faster random number generator. Is it really slower than my old algorithm?

...

You're kidding me.

We all expected many kinds of problems with AIs when then arrived. I don't think laziness was on the list.

From a talk I plan to give tomorrow.

Wonderfully, AI helped me write this slide.

(I rewrote more than half of the slide to make it work better, but that's about par for the course. And yet, still was helpful. :D)

You are correct. I happen to even have a picture of it. And now "tomato sauce" makes more sense as a literal translation. It was all very nice. I was not expecting the strawberries.

In LLDB, I'm having trouble getting symbols for SwiftUI code, even though Xcode can symbolicate the frames. I can see my own symbols in the backtrace, just not the SwiftUI frames. Not sure if this is tied to Emerge's writeup. I'm not trying to symbolicate a crashlog. I'm in Xcode, using lldb. The symbols are right there in the left-hand panel. I just can't find a way to get them in text form. (I need to write scripts that evaluate these traces to isolate a SwiftUI bug.)

https://www.emergetools.com/blog/posts/symbolicating-swiftui-and-any-apple-framework

Emerge Tools Blog | Symbolicating SwiftUI (and any Apple Framework), Part 1

How we managed to symbolicate SwiftUI crash logs.

TSLZ (2X Inverse TSLA) is still a terrible investment if you judge it by how much money it returns. Even after today, I'm down $31. But today felt good. I lost a lot of money today in other stocks. But TSLZ was up 20%, and that felt good.

I'm willing to lose up to $10,000 on this. So I've got a lot more room to keep shorting Elon Musk until he stops shorting us.

Inverse ETFs are terrible investments. I cannot stress enough that you should not follow me with money you need. But if you can…

Spending Imbolc in #NewPort #RhodeIsland for the first time was just the vacation we needed. Folks remarked about our rambling about in sub-freezing weather, but we love to walk on cold, rocky beaches. A beach cabin in winter is the best.

I can’t imagine going in the summer. It’s got to be a madhouse of traffic with those roads. But maybe Oct/Nov when the leaves are turning would be nice. There were a lot of gardens we’d like to have seen. But their trees are amazing in winter.

With all the cool AsyncLineSequence stuff and the interest in "not just apps," I had the feeling that working directly with files was getting better? Is it still kind of a mess? I want to append to a text file, creating it if it doesn't exist. Do I really still need to check with FileManager whether the file exists, using a String rather than an URL, then use four lines of FileHandle code for appending (and separate code for creating)? Are we still not past this?

I hereby absolve you of one piece of your async testing stress.

When you are testing "eventually X will be true," you can poll.

You do not have to create elaborate callback schemes.

You do not have to inject mocks that hide XCTestExpectations.

You do not have setup an AsyncStream or a Notification.

There is nothing wrong with asserting that something will eventually be true. It is exactly what you mean.

And in a unit test, there is nothing wrong with polling for that.

I guess white wine with a bendy straw is a certain vibe. Lakeside Restaurant in Coventry.