fiddlerwoaroof

I do <Esc>k

I’ve used Lazarus a bit and it mostly replicates the APIs and experience of using Delphi in this timeframe

Yeah, my experience is that the rapid growth of computing capabilities of my childhood more or less ended in 2013: the desktop I built then is fine for most of my uses and the significant performance improvements since then haven’t been CPUs getting noticeably faster but increases in the speed of disks (SSDs) and growth of memory sizes and network speeds.

If you have your CI runner use the same commands as local dev, CI basically becomes an integration test for the dev workflow. This also solves the “broken setup instructions” problem.

Imo, the “activation energy” is a lot lower than it appears too. The determinate systems nix installer solves a lot of the local development issues and it’s fairly easy, as a first pass, to write a simple derivation that just copies your current build process and uses nix for dependencies.

Only if you consider connecting to a single phone. The advantage of what I’m proposing is the watch would automatically work with whichever other Apple devices are on your account.

I think I was pretty clear. You setup your pebble watch via openid connect/oauth like any other API client. No nag popups, manual Bluetooth pairing, etc.

Any form of dependency injection by classpath scanning is an antifeature

> This seems awfully close to "avoid Spring" to me, heh.

Exactly :)

I absolutely hate these sorts of nagging popups and I’m happy that a vendor I already have to trust doesn’t pop them up when I acquire a new product and sign it into my Apple account.

Imo, if this were to happen, it should happen by allowing devices like the pebble watch to sign into an Apple account and acquire permissions through that process rather than nagging on my phone on pairing.

Do things like inline method and inline base class work correctly? It’s not a given that IntelliJ is works correctly for Kotlin if the semantics of Kotlin are harder to analyze statically.

Are record patterns stable yet?

I find Java 21 very pleasant, as long as you figure out how to avoid ugly things like Spring’s autowiring of dependencies.

I haven’t really given kotlin a fair try, but I find it ugly the couple times I’ve tried to work with it.

What I’m really curious about: do IntelliJ’s refactorings work as well for kotlin as Java? I find the refactoring tools make up for Java’s shortcomings as a language.

I think the saying is oversimplified (or else, chrome isn’t a product but a byproduct). It makes a company to make the complement of their product cheap or free so that people can use their product more effectively. Google Search and Ads rely on people having a browser. Similarly, Apple’s Phones and laptops are benefited by including a good first party browser. Imo, a lot of technologies are better as free byproducts of monetized technology than as being some company’s products: there are endless lists of great technologies that never found a market or disappeared post-acquisition.

Well, I for one don't want to pay $15/mo for my browser: I'm happy for it to be the complement of some company's product.

There isn’t a business model for a company whose product is just chrome, though. Unless it’s some kind of foundation or non-profit with a sufficient endowment to not need a business model.

I’m no fan of Google, but—unless people really want browser subscriptions—I don’t see how you can have a browser company today. They exist better as a byproduct of some other business model.

Bundles only include things that can be in the pack files: the git objects `git clone` would pull.

I have mixed feelings about discoverability. I'd usually rather functionality be available in a non-discoverable fashion than unavailable because designers couldn't figure out how to surface it in the UI. And, one of the things I like about vim and emacs is the interface can be extremely minimal with all the functionality I need behind key shortcuts and the like. (Although, emacs has a nice solution of `C-x C-b` list all the available shortcuts for the current window and `M-x` listing all the available commands.

The way I learned these shortcuts was reading macOS's help back in 2007, but the quality of documentation bundled with macOS has gone downhill quite a bit.

I never use the taskbar on any system because it’s slower than shortcuts like this.

The command-tab switcher has a lot of hidden functionality:

I always only use cmd-tab to open the switcher, then I use arrow keys to pick an application, up/down arrows to view an application’s windows (arrow keys and enter to select a specific one)

You can also hit Q to quit an application from the switcher and probably more things I’m currently forgetting.

So, the easiest way to swap between windows:

Cmd-Tab, up arrow, arrow keys to the window you want, enter

I empty the trash with command-shift-backspace. I use the dock icon to open it, but I think there’s an entry in the Go menu

I used to hate JVM stacktraces but they’re incredibly useful and I miss them when I’m using other ecosystems these days. The way JVM exceptions do “caused by” when an exception is thrown while handling another exception gives so much more information than other languages. It’s also relatively simple to filter out the noise when you want a concise error message: just remove lines that start with a tab.

CIDER (the emacs IDE for Clojure) does a really good job of filtering the stack trace and presenting several views depending on what you care about. L

The thing is, this is a catch-22 problem: if everyone used the default multiple select, people would know how to operate it (and they’d be able to transfer their knowledge from how the control works elsewhere in their os to the web: for example, if you know how to select multiple files in Finder, Windows Explorer or in a open file dialog, it works exactly the same way in the multiple select control). But, because web developers have decided to invent a 1,001 different multiple selects, the default control is inaccessible.

The real issue I have with the web is everyone has broken user interface consistency across the platform with custom web controls where the default controls would be fine.

I don’t understand the hatred for select multiple on desktop. It’s a list box that uses standard UI conventions that have been present for three or four decades now.

The issue I had when I still worked in the frontend world is the frequency of major version updates. If you used standard project bootstrapping tools in approximately 2017, you got a whole bunch of libraries and tools that would have significant breaking changes multiple times over the next three years or so (I know because I had to maintain them).

This isn't true: local (to me) jurisdictions vary between an 8ish% sales tax and a 10% sales tax. In many states, there's no sales tax on food but food is taxed in Alabama. There's any number of variations of this in the US.

One thing I like about the nix language is it uses semicolons to separate elements of the map: while I use trailing commas, they always look dangling to me whereas semicolons look fine without another expression following.

Good to know, I passed over Ada for Common Lisp around 2015 because of this.

The binaries AdaCore used to distribute stated that you had to get gnat elsewhere for this link level exception because the runtime library they distributed did not have the exception. But, I'm glad to know that this restriction has been solved.