Lars Fosdal

Not sluggish... FUBARed. Why can't I cancel the deployment phase? Why does the IDE lock up after deployment?

is and as could help declutter that versatility, though.

The appropriate place would be something like \Program Files (x86)\<Your Company or System Name>\<Services>\YourServiceHere.exe

The settings look very different on Windows 11. Some of them are not enabled for RDP.

Judging from what I read, it seems that x86 runs fine, but x64 is still work in progress on MS side. W11 for ARM still is in preview state. I'll be able to answer this for myself, around Xmas 🙂

On the topic of VMs - I wonder how well Windows 11 for ARM will be running under Parallels and if it can run x86 and x64 code.

Hyper-V.

Yes, it shows. That thing is f'ing brilliant!

I like 11. It appears to be snappier than 10. But, yeah... there are still remnants from 7 in there. Really annoying that they haven't done away with the old Control Panel completely and got it modernized in Settings.

Where do you get T2 from in the outer EXISTS?

COBOL is pretty verbose...

Only the start menu has rounded corners. Go figure.

Annoyingly, the Windows 11 GUI doesn't show rounded corners in my VM. Makes me wonder what else looks different between a VM and a regular desktop.

We use inline vars quite a bit. For loops, in particular - as well as for temporary variables. But, yes, the tooling is lagging behind.

You've put good advice in that article, @Uwe Raabe. Just my two cents... A lot of people say multi-threading is hard. It is only hard if you have not learned to do it properly, like any other thing in life. Debugging multi-threading apps, on the other hand... Make sure you log a lot - with the thread context in the log. IMO, the debugger is not yet properly equipped to deal with multiple threads, unless you count using lots and lots of breakpoint tweaking. The Parnassus Parallel Debugger plugin may solve some of that, but it is not yet out for Delphi 11, and it has stability issues in 10.4.

FWIW: We have several 32-bit services. There is no reason what so ever to put them in one of the Systems folders.

Just an observation. I've been on Windows 10 for a few years now, and I don't recognize this "Windows keeps downloading stuff". Once a month, there are patches - and staying patched is generally a good idea. As long as you are on Windows Pro or Enterprise, you have control over WHEN you want the patching to happen, unless your company has overriding policies - in which case you need to butt heads with the IT department to get them adapted to your needs. If you are using Windows Home, you need an upgrade. Disabling/Enabling your network connection on the PC to avoid patching seems like the wrong solution, because you really need that working connection to get work done. If you really need to block something, do it in a firewall rule. Most routers can do that.

In concept, it is great - but lack of support in the debugger is a problem. If you have two blocks with var x := <whatever> it gets even more confused. I love using inline vars for loops -> no use after the loop block and type inference saves you from having to declare the type.

When you look in the form designer, are the OK and Cancel buttons correctly parented? I.e. are they on the expected place in the form hierarchy?

It is VERY hard. Especially when you can't persuade the debugger to break ONLY in the thread(s) that you want to debug.

I am about to embark on learning IBM Cloud Pak for Integration - which is a low code solution for API development. The low code bit is a buzzword, because you are looking at learning and applying a lot of ESQL to achieve what you want. It feels like early VB / Delphi in a way - only with a worse UI.

In that case, I would have created a worker thread (or threads) that have a private RIO, work in, work out queues, and a default heartbeat ping loop method that is called when the work in queue is empty .

I am curious - why would you not create it when you need it? I checked our code, and we typically create on demand per thread.

When it fails, does it always fail (on that machine)? Is there any red thread such as OS version and patch level, display driver, etc between the machines that fail?

As long as the RIO is used only by one thread at a time, that would work fine.