Lars Fosdal

If you expand it with the examples that actually works - it would be sufficient for a QP, IMO.

Interestingly, the two following changes eliminates the AV on raise. Either, change LogException to a procedure procedure LogException(const E: Exception); begin Writeln('Log + ' + E.Message); //Logging('Blah blah' + Result, 'Exceptions'); FYI end; or assign the result to a variable: on E: Exception do begin lMessage := LogException(E); // AV in Sydney, OK i Berlin // LogExceptionEx(E, lMessage); // OK raise; end; It does look like a compiler problem. Is there a QP report?

Clearly, it is based on jokes alone.

I guess injecting a deprecated statement for the old class could be an option - or is that too brutal.

@Jim McKeeth should kick some shins to have a new roadmap produced.

if WinApi.Windows.IsIconic(Application.MainForm.Handle) then begin // Log something end; https://docs.microsoft.com/en-gb/windows/win32/api/winuser/nf-winuser-isiconic?redirectedfrom=MSDN

I can't remember why not - prolly because Numlock was the first to spring to mind. Anyways, I've been using this for years now, and never experienced any adverse side effects.

Have any of you seen something similar to this? Delphi 10.3.3

Well, it sort of forced my hand in switching to 10.4.1 a little ahead of time. Going live on new version to day. Oh, such fun to reconfigure all the build scripts 😛

Sounds like an excellent suggestion!

Much annoying. Very bug. 😛 https://quality.embarcadero.com/browse/RSP-30576

Doh! Ref. https://blogs.embarcadero.com/ide-plugins-in-rad-studio-10-4-1/ The solution was to use GetIt to uninstall 1.6.1 and install 1.6.2.

There is an extra fee for solving homework problems, isn't there?

Thread split off from if Data.TryGetValue(aId, vNewData) then vNewData.DataName := aName; else begin vNewData := TDataRec.Create; vNewData.DataId := aId; vNewData.DataName := aName; Data.Add(aId, vNewData); end; @Stefan Glienke - What is the best container for random lookup access for records?

Indexing and doing index maintenance is crucial to maintain performance. If the queries can be done completely free form, that will be a challenge. IMO, there are few settings to FireDAC that really affects the performance once the query has been passed to the server.

In my experience, it is far easier to use classes when you need to modify contents. I wish it weren't so, and it can be worked around - but then usually with pointer references, and at that point (pun intended), I might as well be doing objects and maintain type safety and support polymorphism.

Can you offer an example?

I am trying to improve my tendency to spin off a discussion in a thread focused on something else, so here we are. To reiterate, my preference for classes over records with regards to generics, is that read-only generics for records works perfectly for its use, but when you need to update the content in the record, it becomes a very different exercise. You either need to have structure specific methods to change the record field values, or you must copy the entire record, modify it, and the copy it back. The first option wreaks havoc on reuse, and the second option can become a performance bottleneck if the record is large. Even more so, if the record is multidimensional, containing records with arrays of records. The same kind of change in a multidimensional class, is trivial in comparison. @David Heffernan - hence my "the amount of reuse" argument.

Moved my digressions to

"the amount of reuse" was my point.

We also know that the two need to be handled differently when it comes to changing their contents.

I splitted this thread off this

Splitted off the Interposer discussion.

If I did the my own component approach - I would need multiple versions of the design time components to deal with the different methods of feeding data to the controls, and every time I upgraded TMS, or made a change to the wrapper, I would have to rebuild them. Since the wrapper code is continuously evolving - switching between Live, Pilot and Trunk would be very cumbersome if I had to swap the components used as well. Since the data side of the wrapper is not linked to visual code and the events are connected to the wrapper and not the component, it is easy to write non-visual test code that can simulate UI events. Custom components are an option, but. I personally prefer the wrapper approach.

We use interposers or wrapper classes to populate third party components such as the grid, list and combo controls from TMS. It started with us having hundreds of grids, and to get a normalized look, it was better to use a wrapper class to set them up, than to manually configure each class in each form. This class eventually evolved into several wrappers, most of them being data-aware so that we had unified ways of populating a grid, list or combo - straight from SQL, from object lists, or direct fill from "hard code", as well as handling selection events, etc. Another benefit is that if TMS should happen to make changes or add features - we only need to deal with it in the wrappers.