Robert Ryan

3 Followers
14 Following
28 Posts

@cocoaphony Yeah, it's a new issue. I don’t know if it is my transition to Apple silicon or macOS 13.3.1 or some recent Charles update or what have you. They all happened over the last few months and I don’t use Charles often enough to diagnose the precise moment it wigged out on me.

I had the isolated Charles-induced problem in the past, but they were edge cases. What I'm now seeing is a general flakiness in the networking stack after I do a lot of Charles debugging. And problems persist even after I quit Charles.

I’m just trying to figure out if it’s just me, or if others have the problem, too.

#CharlesProxy #FlakyNetworkStack

Any macOS users having troubles with Charles HTTP proxy messing up their network configuration? If you're a macOS Charles user, let me know, either way.

For me, it works for a while, but I eventually find my network getting hosed, only fixed with a reboot. I don't like the Wireshark UI nearly as much, but Charles is just not worth the headaches.

(Running macOS 13.3.1 on M1 Mac; Charles Proxy 4.6.4.)

In 2002’s *Minority Report*, they kind of destroyed the “halo” product name for me, as it was a dystopian form of restraint. I can see why the Halo Collar (https://www.halocollar.com) got away with it, but it always struck me as a very odd name for Amazon’s fitness/health products. I guess that problem now goes away.
GPS Wireless Dog Fence | Halo Collar

Explore Halo GPS Wireless Virtual Dog Fence with enhanced features, like tracking your furry friend's activity for ultimate safety, even on-the-go

Halo Collar
I find it curious that `@objc` methods are not inherently `nonisolated`. What does it mean to say “this is available to Objective-C”, but also is a member of the actor isolated type?
Now that I fully understand the new pricing scheme, I am no longer conflicted: I am jettisoning Postman. Asking us to pay $1200/user/year to run tests on our own hardware is just outrageous.
@cocoaphony IMHO, you almost ways always do it precisely as ChatGPT explained, because inside the expiration handler, you always check to see if it is `.invalid` and if not, cancel it and set it to `.invalid` (resolving potential race with your asynchronous process). But you can’t refer to `backgroundTask` inside the expiration closure in the same line in which you instantiate it. The pattern in the ChatGPT answer is right (though they made it more confusing than necessary because it didn't show what is going on inside that expiration handler).
When ChatGPT answers are pointed out, the offender generally quickly deletes the answer, both recovering downvote in rep and hiding the plagiarism. Maybe I’m going off the deep-end here, but maybe deleted/flagged answers should be visible in user’s history…
OK, I’ll bite: What is weird about the `var`? The last thing I want to do is defend ChatGPT, but if you only need a few seconds to finish a network request after the user leaves the app, this is the right way to do that…

I am conflicted re Postman evolution: I really appreciate the recent UI enhancements. But, I am not crazy with how they handled the lowering of the freemium bar (now only 25 test runs per month, even when doing this locally!). Worse, there was no clear warning, so I just got cut off. So, I’m reverting to CLI for now. The GUI is nice, but not enough to justify the recurring subscription price.

As an aside, I'm far more open to subscription pricing when I’m using ongoing cloud infrastructure, but I *really* don’t want to use their cloud services. (Historically, their cloud integration has introduced a *lot* more problems than is justified by the meager value prop.)

Hey, I completely get the need for subscriptions for sustainable biz model, but pricing for doing everything on my boxes should be different (and lower) than one that uses their cloud infrastructure. The price, for indie developer, is not outrageous, but it’s high enough that I have downloaded and am testing alternatives.

I am seeing enough folks trying to pass ChatGPT answers off as their own original content, that I have to ask *why* they’d bother doing so. The only think I can think is that they want to have future employers assess them on the basis of their postings. But they’re oblivious to the fact that it’s generally pretty easy to sniff out ChatGPT content (and is to their detriment when uncovered). As an employer of IT talent, it’s certainly been added to my list of things to check.