eivindbakkestuen

You are describing a wizard UI...

You are doing R&D. Are you saying your R&D process does not allow you to use an exception stack trace logging solution to figure out where it happens?

Have you looked at the Message parameter values to see what you're actually getting?

We are aware that the UI isn't modern anymore, and work is underway to improve that. However, @Tom F is right to say it just works (up to and including the latest Rio version!). Which allows the users to get their work done and move on. 🙂

Out of idle curiosity, does this happen no matter how many cores you have in the CPU? What do the CPU core graphs look like in the Task Manager?

Have a look at this which was just posted: UberFrame panels

You may want to test docking. You're not locked to docking the windows in particular places, they can float around on the "desktop" inside your application. You decide what kind of docking etc operations are applicable (perhaps you have a few child windows where it would be appropriate to only see one of them at a time, so would work in a tabbed child window, etc). I'm attaching an image of an application using the DevEx docking library; could just as well have been done with MDI.

Are you still using version 10.1? The issue was apparently fixed in 10.3.1, perhaps test with the Community version to make sure before you go for the upgrade.

Ah, your BoundIP reference got me thinking. After some sleuthing, It turns out that my client had been misinforming me; BoundIP was put into use some months ago, and I was told that version worked for them. In reality, they were still running an older version which did not set BoundIP, which then caused confusion when a new update was delivered and this error immediately triggered. Lesson: don't trust clients (check), always use exception stack tracing (check) and log everything (check).

The thread (and the TIdHTTP instance inside it) are destroyed as soon as the result of the Get() call is available. The number of threads would not have been great (less than 10) since this happened shortly after restarting the application. This exception never happened in the previous 3 months of intensive application use, hence I'm puzzled. 🙂

Here's an unusual one: I'm using TIdHttp.Get(<IP:port of local network machine>\somepath) inside a short lived thread (if that matters); no specific ports or anything is set in the IdHttp instance before the call. The other day, we got a number of exceptions like below on this call. I'm just wondering if anybody else has seen it, or has an idea what may have caused it? Not intending to blame Indy here, at a guess the OS doesn't reuse ports quick enough (limited resource) or something of that nature. However as part of a customer critical system, it'd be good to know if there's anything that can be done about it? (forgot to say, the application runs many instances on a Terminal Server) Exception: EIdCouldNotBindSocket Message: Could not bind socket. Address and port are already in use.

Just grab NexusDB Embedded Free from GetIt, compiles directly into your windows app, only your exe to distribute.

Er... the OmniThreadLibrary zip is only 8.8Mb, so... seriously... 😉

The latest dev snapshot comes with batch files for Delphi 2007 and older, but the packages can't be compiled (first problem, stops on UInt8 etc types in idBuffer.pas which doesn't exist in D2007). Is D2007 meant to be supported still? I couldn't find a definitive document to say what IDE/compilers are currently supported.

In this primary English forum, it'd be nice to have the translation of those messages... 🙂