Willicious

The image that's being clicked is to show the modal (Level Select) in the first place, which is where the treeview is displayed. The call stack doesn't show the treeview click event because I've inserted a breakpoint at the treeview click even handler call, and I'm pausing the program between clicking the treeview and the breakpoint happening - so, we never see what the treeview is calling in the call stack. That's OK, though; the point of this exercise is to determine what's causing the delay between treeview click and treeview click event handling. Regarding the ApplicationIdle procedure, I find it unlikely that this is the problem. As I stated in my previous post (important parts highlighted): So, hopefully we can agree that the ApplicationIdle procedure is unlikely (although not uncertain) to be the problem. To address your questions, though: I didn't write the code, I'm working on a fork of no fewer than 4 other programmers' work. So, the codebase is full of inconsistencies and calls that seemingly have no purpose, etc. As a relative beginner (I've only been working on it for about a year and a half), it makes the process that bit more difficult, but I do get to learn lots about how not to do things. The ApplicationIdle procedure is intended to do a few initial checks upon first loading the program (is it a new version?, does a user config file exist?), and then handle the situation accordingly. Generally this is only relevant the very first time a brand new version is opened on a user's system. Beyond that, it handles updating the scroller reel (a simple animation on the menu screen). My best guess is that the sleep allows time for the reel to update, but I could be wrong about this. The original developer is difficult to get hold of, and usually can't remember off the top of their head why certain calls have been made, so I have to make (un!)educated guesses and generally leave things as they are unless I have a good reason to change them. I didn't know I had to, simple as that really. Best guess is that some previous version of the unit did set Done to true, but this was later removed (perhaps inadvertently). I'm unsure of exactly what to do about this, would you suggest I simply set remove the call to set it to false, or explicitly set it to true somewhere else (possibly when closing the menu screen)?

Commented out the entire ApplicationIdle procedure (including the call for Sleep(1)). The result is that the scroller animation in the Menu screen doesn't update. Otherwise, the program runs normally. However, the problem (i.e. slow-responding Level Select treeview) persists. Here's the resulting call stack (i.e. between physical mouse click and treeview click event handler): Here's the ApplicationIdle procedure: procedure TGameMenuScreen.ApplicationIdle(Sender: TObject; var Done: Boolean); begin if fCleanInstallFail then begin DisableIdle; fCleanInstallFail := false; ShowMessage('It appears you have installed this version of SuperLemmix over ' + 'an older major version. It is recommended that you perform a ' + 'clean install of SuperLemmix whenever updating to a new major ' + 'version. If you encounter any bugs, especially relating to ' + 'styles, please test with a fresh install before reporting them.'); fLastReelUpdateTickCount := GetTickCount64; EnableIdle; end else if not GameParams.LoadedConfig then begin DisableIdle; GameParams.LoadedConfig := true; ShowSetupMenu; EnableIdle; end else if not fDisableScroller then UpdateReel; Done := false; Sleep(1); end; Any suggestions welcome...

I'll definitely give it another try at some point. For now I'm trying to get a few things stable before I introduce any new project elements (hence trying to get the treeview code optimised).

I see. I'm guessing there's no easy answer to this particular problem. Could I maybe limit the number of messages being sent in some way, or prioritize treeview stuff when the form is active? Managed to grab the Pause button from the Menu just in time. Anyone's guess why there's no hotkey for this; someone at Embarcadero didn't have their Weetabix the day that feature was added, clearly. Anyways, here's a screen shot of what I can see. CPU processing and no call stack history: However, did get this when I stepped over with F8: Very unlikely. It's a 32-bit program which (as far I understand) doesn't support multi-threading. It's pretty ancient by today's standards so unlikely to be trying to pull those sorts of manoeuvres (although, I could be wrong about that). This is more likely. Caching sounds like a very good idea, since each time a treeview item is clicked it has to reload a bunch of info. As far as I can tell, all assets are loaded on form creation, but there could be some re-loading/re-drawing going on somewhere. Thing is though, if the delay is happening before SetInfo is even being called, then it stands to reason that isn't the cause of it. Still, I'd definitely be up for learning how to cache stuff if it's a generally helpful thing to do to save memory. Yes. I've had someone else suggest that I use a better treeview. I'm guessing I'd install it like a plugin, like Bass.dll or GR32 (or madExcept)...? Would definitely be up for exploring that. See above screenshot for debug output on pausing. I'm probably not using the debugger correctly though, because it's gone straight to CPU stuff. Meanwhile, madExcept found nothing, so I've uninstalled. It probably is the message queue, from what's being said. --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Thanks for the help so far everyone, I know I'm a bit of a novice; it's nice to have the suggestions.

Last resort as it takes a while to paste in debug strings to every procedure, but if all else fails this is probably what I'll do In fact, calling it now that this is probably what I'll end up having to do

This one is very useful to know in general, thanks. Does Program Pause have a hotkey? Since I'm needing to debug input, I need to be able to pause pretty much immediately after I've clicked the treeview.

This helped, thanks. I put a breakpoint at the only procedure called by the treeview's click event handler: procedure TFLevelSelect.tvLevelSelectClick(Sender: TObject); begin SetInfo; <-------------------- breakpoint here end; What I noticed is that there was occasionally a significant delay between clicking the treeview and the debugger pausing the program, but only for certain treeview items. Some pause immediately, others take longer. This is consistent with the behaviour I'm noticing with the release version. I'm interested to know what this might point to. Why would the program sometimes be taking a while to acknowledge input?

Thanks for the suggestion! Installed, enabled, compiled. It almost seems a bit pathetic to be typing "what do I do now?" but... I really have no other option Maybe I can buff it up with some other more specific questions: Does madExcept have a GUI? If not, how can I see its output? Can I point madExcept to a specific .pas file to look for possible optimizations? If not, how can I know that madExcept is looking at what I need it to look at? You know, that sort of thing EDIT: Better question - how would I know if the hangs are caused by the app pumping the message queue?

The app I'm working on has different "screens", which are each essentially large bitmaps. When closing a screen to go to the next one, the following code provides a fade-to-black transition effect: procedure TGameBaseScreen.FadeOut; var Steps: Cardinal; i: Integer; P: PColor32; StartTickCount: Cardinal; IterationDiff: Integer; RGBDiff: Integer; const TOTAL_STEPS = 32; STEP_DELAY = 12; begin Steps := 0; StartTickCount := GetTickCount; while Steps < TOTAL_STEPS do begin IterationDiff := ((GetTickCount - StartTickCount) div STEP_DELAY) - Steps; if IterationDiff = 0 then Continue; RGBDiff := IterationDiff * 8; with ScreenImg.Bitmap do begin P := PixelPtr[0, 0]; for i := 0 to Width * Height - 1 do begin with TColor32Entry(P^) do begin if R > RGBDiff then Dec(R, RGBDiff) else R := 0; if G > RGBDiff then Dec(G, RGBDiff) else G := 0; if B > RGBDiff then Dec(B, RGBDiff) else B := 0; end; Inc(P); end; end; Inc(Steps, IterationDiff); ScreenImg.Bitmap.Changed; Changed; Update; end; end; Is there any way this same code can be modified to create a new FadeIn; procedure which allows the screen images to fade-in-from-black? I'm guessing I need to first specify that RGB := 0, 0, 0 and then increase the RGB values until they reach the actual bitmap values. Do I first need to "get" those values somehow? Or, is there a better way to achieve the Fade In/Out procedures? (note that we're dealing with combined bitmaps here, rather than single images)

@Anders - Brilliant! This last one worked a treat. I copied in the ShowScreen procedure as an override and it works great for all screens except the play screen (in this case, it fades from white instead of black). However, I also noticed that the latest version of the FadeOut procedure (the one from the PR) also fades to white on the play screen. So, I'm guessing that this is because FadeIn and FadeOut use MasterAlpha, and somewhere in the Lemmix codebase the background of the play screen is set to white, but I could be wrong. I'll look into this more tomorrow. Thanks again, it's taken a long time to get anywhere with this!

OK, I think the reason why none of the attempts at a FadeIn procedure are working is because the UI isn't actually responding to the updates. This could either be because the App is idle at the point the FadeIn is being called, or because it's only processing the first message from the FadeIn procedure, and then none of the others. At one point, I tried displaying an animated "Fade" graphic (basically a large black square which gradually faded to transparent over 8 frames) over the existing screen image - I could only ever get it to display one frame of this animation. The app is definitely capable of processing messages and animations in the menu screens, though, because other animated images (a text scroller) are shown, clickable regions have graphical effects applied to them, and the FadeOut procedure works. Here's where FadeOut is called, maybe there's a clue in here somewhere: procedure TGameBaseScreen.CloseScreen(aNextScreen: TGameScreenType); begin Self.OnKeyDown := nil; Self.OnKeyPress := nil; Self.OnClick := nil; Self.OnMouseDown := nil; Self.OnMouseMove := nil; ScreenImg.OnMouseDown := nil; ScreenImg.OnMouseMove := nil; Application.OnIdle := nil; fScreenIsClosing := True; if fCloseDelay > 0 then begin Update; Sleep(fCloseDelay); end; FadeOut; if GameParams <> nil then begin GameParams.NextScreen := aNextScreen; GameParams.MainForm.Cursor := crNone; end; Close; PostMessage(MainFormHandle, LM_NEXT, 0, 0); end; And, here's where FadeIn is currently being called: constructor TGameBaseMenuScreen.Create(aOwner: TComponent); begin inherited; fKeyStates := TDictionary<Word, UInt64>.Create; fClickableRegions := TObjectList<TClickableRegion>.Create; fMenuFont := TMenuFont.Create; fMenuFont.Load; fBasicCursor := TNLCursor.Create(Min(Screen.Width div 320, Screen.Height div 200) + EXTRA_ZOOM_LEVELS); LoadBasicCursor('amiga.png'); SetBasicCursor; InitializeImage; OnKeyDown := Form_KeyDown; OnKeyUp := Form_KeyUp; OnMouseDown := Form_MouseDown; OnMouseMove := Form_MouseMove; ScreenImg.OnMouseDown := Img_MouseDown; ScreenImg.OnMouseMove := Img_MouseMove; fCalledFromClassicModeButton := False; FadeIn; end;

Thanks again. Made the change, still just getting a black screen. I'll have more time to look over it tomorrow properly, it may be because of where I'm calling the procedure from or something like that.

Thanks for the code snippet Anders, I've given it a quick try and just get a black screen. I'll investigate further later this evening or tomorrow, what you've written looks like it should work. I'll do some debugging and see if I can figure out what's going wrong.

Tried this again today. This yields the same result as always (black screen for 1000ms, then the screenimg.bitmap is drawn at full opacity). I've tried placing "Inc(i)" at various points in the loop, and have also tried using "repeat ... until i = 255" to no avail: procedure TGameBaseScreen.FadeIn; var i: Integer; RemainingTime: Integer; OldRemainingTime: Integer; EndTickCount: Cardinal; const MAX_TIME = 1000; // mS begin EndTickCount := GetTickCount + MAX_TIME; OldRemainingTime := 0; RemainingTime := MAX_TIME; ScreenImg.Bitmap.DrawMode := dmBlend; // So MasterAlpha is used to draw the bitmap while (RemainingTime >= 0) do begin if (RemainingTime <> OldRemainingTime) then begin for i := 0 to 255 do begin ScreenImg.Bitmap.MasterAlpha := i; ScreenImg.Update; end; OldRemainingTime := RemainingTime; end else Sleep(1); RemainingTime := EndTickCount - GetTickCount; Inc(i); end; Application.ProcessMessages; end; For reference, here's the FadeOut procedure as written by Anders, which works perfectly: procedure TGameBaseScreen.FadeOut; var RemainingTime: integer; OldRemainingTime: integer; EndTickCount: Cardinal; const MAX_TIME = 300; // mS begin EndTickCount := GetTickCount + MAX_TIME; OldRemainingTime := 0; RemainingTime := MAX_TIME; ScreenImg.Bitmap.DrawMode := dmBlend; // So MasterAlpha is used to draw the bitmap while (RemainingTime >= 0) do begin if (RemainingTime <> OldRemainingTime) then begin ScreenImg.Bitmap.MasterAlpha := MulDiv(255, RemainingTime, MAX_TIME); ScreenImg.Update; OldRemainingTime := RemainingTime; end else Sleep(1); RemainingTime := EndTickCount - GetTickCount; end; Application.ProcessMessages; end; Any suggestions or comments would be most welcome!

It's in the title. My current CE licence expires tomorrow, and all efforts I've made so far to obtain a new licence have failed. Does anyone know how I can obtain another 1-year licence for this particular edition of RAD (Sydney 10.4.2)? 10.4.2 is the only version that will successfully compile the project I'm working on, and I need to be able to seamlessly continue with it if at all possible. Both 11 and 12 run into problems when getting the project set up, so simply installing the latest version unfortunately isn't the solution. Please help!

I'm a hobbyist and can't justify paying for a Professional version given my limited use of the platform. Also, as stated in the OP, my project only works in 10.4.2 so even if someone were to gift me the latest Pro version, it would be useless to me. Does anyone know of a way to get a new serial for 10.4.2 CE? Anyone had any luck with contacting Embarcadero for help?

Apologies, I didn't explain that properly. I'm not looking for a single point to be true, but rather a number of points along a straight line. So (X = 0), (Y = 0 and 1 and 2 and 3, etc...) would draw the line required. But, the entire line needs to be drawn for the condition to be met. Yes, to be honest I do now remember asking this same question a while back. Here it is. I didn't fully understand the answer given in that topic, but now that I look at it again I can see what's being suggested. It seems that a boolean needs to be set, then each number is iterated over until one returns false (which breaks the loop). If all numbers return true, then we can run the code. That could work, and using "case ... of" could also work, again to set the bool value. Thanks everyone, I'm sure I'll figure it out.

I need to use a full set of numbers as a condition. So, for example, rather than: var i: Integer; for i := 0 to 5 do begin if PosY = Y - i then //whatever end; I need it to be: var i: {not sure what to declare i as}; i := [0..5]; if PoxY := Y - i then begin //whatever end; The code needs to execute iff all conditions are met (Y - 0 and Y - 1 and Y - 2 and Y - 3, etc) and not just if any of the possible 6 conditions are met (Y - 0 or Y - 1 or Y - 2 or Y - 3, etc). I've tried declaring i as "array of Integer;" but get the error "Incompatible types: dynamic array and set". I can't declare it as a set either, so, what do I need to declare i as in order to be able to specify all values between 0 and 5 as the condition?

Yes, this is what I'm after. Thanks! 🙂👍

Apologies, I appreciate the suggestion and the time you've taken to respond, but this isn't quite what I'm looking for. I need a program that will basically "read" the entire codebase - ideally at source or within the IDE itself - and point out any areas which could be removed, simplified, or optimized to reduce performance bottlenecks. Maybe I'm dreaming and such a thing doesn't exist yet...? I just know that AQTime used to be bundled with RAD Studio but isn't any longer (and, from what I've seen, this does exactly what I'm describing, but is currently mad expensive for someone who only does small open-source projects). Also - I'm looking for somthing that will do the above, and which works straightaway with an absolute minimum of setup. If that doesn't exist, then that's fine. Maybe it will someday 😊

Hi all, I'm looking to find a decent, easy-to-use Profiler to help identify performance bottlenecks and memory leaks. I'm a small-project user so not looking to pay too much - free and open source options are welcome as long as they're straightforward to use - no command line stuff, please. It's 2023 and I'm using RAD Studio 10.4.2 - I believe SmartBear used to provide a standard version of AQTime with an earlier version of RAD, shame they stopped doing this really because now it seems we have to pay big bucks for this particular profiling system. Recommendations and suggestions welcome, thank you!

I take it VTune is the Profiler... what is the MAP2PDB for? Complete beginner here, I've been advised to Profile my project's codebase to look for permance issues. Not sure where to start. Also, isn't VTune an expensive Intel app? Ideally looking for free, or inexpensive alternatives. VTune installer says it can integrate with MS Visual Studio, but not RAD Studio - is this even a Delphi app?

Thanks! I've replied in-topic in case anyone's watching the topic and has any suggestions.

Replying again because I forgot to check "notify me of replies", and the site doesn't provide a way to do this after replying

Hi, looking for a profiler in 2023. I'm a small-project user and so don't really want to pay. I use RAD Studio 10.4.2, any suggestions?