Uwe Raabe

Looking at the current VCL implementation it would be little work to show the question mark icon for mtConfirmation. The current behavior is a deliberate decision and probably targeted to match the Windows design. Imagine a user having different icons for confirmation messages depending on the IDE used to compile the program. So I think the message anyone should ask itself: Do I want a look and feel different to the standard Windows one?

That is not a sign of a dead product. Sometimes bugs are hard to reproduce, sometimes they are hard to fix or simply not relevant enough to be tackled. Don't forget that this is a free product. The maintainer may not have the time or will to work on it as if it were a well-paid full-time job.

In the later versions Delphi checks for Windows 8+ and uses the MS approach. So the docwiki is outdated and has to be corrected.

Are you able to run this reliably in a recent Windows? My attempts showed that it always just returns false.

Seems that Microsoft has just closed that loophole.

It is even documented:

What about TCheckListBox?

It happens in Delphi 10.3 and above. I suspect something in the VCL style drawing of the hint text.

Indeed, I can see the same. It seems like the DT_WORDBREAK is not working correctly.

Sorry for sounding blatant, but it seems you are doing something you shouldn't do. Unfortunately we can't see what it is.

Could it be that you have PrimaryForm.Visible = True in the Object Inspector?

I thought about that first, but IMHO it would be less flexible. Imaging you have some procedures wrapped in an {$IFDEF Debug}. Wouldn't you expect to see those even when the RELEASE configuration is active? With the two options described above, you have control over what the MMX parser sees on the unit level (option 1) as well as on the project level (option 2).

The painting is almost completely done inside VirtualTreeView. I have to investigate why it doesn't pick up the correct style colors.

Yesterday I managed to release V15.1 of MMX Code Explorer. Besides some bug fixes it implements a couple of feature requests from MMX users. The parser system got some significant overhaul, albeit it is still far from perfect. At least there will be less errors in the MMX Explorer tree as well as better support for generics and inline var/const. I rated the improvements worth a bump in the Minor Version number.

There used to be an API for it, but it is no longer available since it has been acquired by Embarcadero. Even with that it would be some work to get this working. If I were you I wouldn't my breath.

What do you mean by work with Bookmarks exactly? Does it crash or something?

MMX V15.1.1 is imminent. I hope the incorporated changes will handle even that scenario.

I am aware of that problem, but did not found a solution yet. MMX gets notified when a package is loaded and triggers an async scan for included components. During IDE startup a lot of packages are loaded and each triggers a scan. There are means to skip a scan when another is triggered, but that is not implemented very effective. So when a lot of packages are loaded sequentially, it will probably led to a lot of unnecessary scans. It would be best when the IDE would notify that all packages are loaded, like it does in the splash screen. Unfortunately there is no ToolsAPI support for that. I am planning to reduce the workload with a more sophisticated approach in one of the future updates. Thanks to the caching of these scan results the following IDE starts are usually much smoother.

I guess, I found a possible scenario where this could happen. It will be addressed in the next version.

Well, actually EurekaLog is not my component. Unfortunately, I cannot reproduce it here, because I don't own a license.

The complete results can be found here: https://survey.stackoverflow.co/2022/ Some Delphi related excerpts: In the Programming Languages category Delphi is listed 7th In the Integrated Development Environment Delphi is listed 3rd The Median Yearly Salary raised from $46,704 in 2021 to $63,984 in 2022

Oh, yes! There is Big Indian and Little Indian format, eh?

Just for clarification: Do you want to hide/show t design time all (usually visible) controls where the Visible property is false?

Probably more something like an incomplete spec.

The IDE places the uses clause after the CLASSGROUP entry when using Alt-F11. While the compiler may ignore it, it seems to matter for the IDE.