bugfix(logic): Fix broken camera edge scroll after loading a savegame from the ingame menu

[stephanmeesters]

Closes #1922

The capture state of TheMouse includes a paused state, however this paused state is not currently reset when a new game is started. This PR addresses this by resetting the paused state in GameLogic:reset().

[xezon]

The observation makes sense.

The implementation is not optimal. It will be better to replace

	m_pauseFrame = 0;
	m_gamePaused = FALSE;
	m_pauseSound = FALSE;
	m_pauseMusic = FALSE;
	m_pauseInput = FALSE;
	m_inputEnabledMemory = TRUE;
	m_mouseVisibleMemory = TRUE;
	m_logicTimeScaleEnabledMemory = FALSE;

with

	setGamePaused(FALSE);
	m_pauseFrame = 0;
	m_inputEnabledMemory = TRUE;
	m_mouseVisibleMemory = TRUE;
	m_logicTimeScaleEnabledMemory = FALSE;

higher up in this function and probably also in the init function.

This way other events can also properly update on pause state change.

Not sure if

	m_inputEnabledMemory = TRUE;
	m_mouseVisibleMemory = TRUE;
	m_logicTimeScaleEnabledMemory = FALSE;

still need to be reset then. I suspect it does not matter, but not sure.

[stephanmeesters]

I've made the requested changes and moved the memory resets above the setGamePaused(FALSE) so that things like TheMouse->setVisibility are updated as expected

Mouse edge scroll behavior, music and sound effects and mouse appear fine after the change. I do now notice a very short audio music glitch when loading the savegame probably (confirmed) because we're now calling TheAudio->resumeAudio(AudioAffect_Music) in reset()

[L3-M]

What is the current status of this change?
Successfully merging this pull request will not close the linked issues ( #1922 and #1999 ) with this current description. Please check GitHub Docs:

https://docs.github.com/en/issues/tracking-your-work-with-issues/using-issues/linking-a-pull-request-to-an-issue#linking-a-pull-request-to-an-issue-using-a-keyword

[stephanmeesters]

The description has been updated to link the issues.

I've been looking for a clean solution, the problem right now is that setGamePaused does not only affect state but also performs actions like resuming sounds and music, which now leads to an audio glitch when you do that in reset() or init(), which ideally is a function only concerned about state.

Also setGamePaused deals with memory variables for restoring variables which does not apply for init() and reset().

So I think either we add some parameter to setGamePaused to indicate we only want state changes and no side effects, or we keep reset() much like it was before and i will do another pass to see that we properly reset everything.

It was possible to clean up init() a bit by using reset() there see refactor commit.

[xezon]

Better make the deduplication a separate refactor change to reduce the complexity of this bug fix.

As for music bug, check why it happens and try to figure out an elegant way to avoid this.

[stephanmeesters]

Alright the deduplication commit is gone from this and I've added a parameter to setGamePaused to prevent resuming audio in init() and reset().

The audio mishap was because reset would resume the music, and then milliseconds later the audiomanager subsystem would also reset which does a hard stop on all sounds

[xezon]

I am confused by this change. The title says that this concerns camera edge scrolling, but the code touches resume audio code. How does this relate?

[stephanmeesters]

In your first review you suggested using setGamePaused to reset the mouse, but this had the side-effect of resuming audio at a point of time you don't want it, so the extra code was added to avoid creating an audio bug.

In hindsight calling setGamePaused in reset() may not be so elegant so let me know if you want a simpler solution that doesn't touch audio

[xezon]

If we can not need Bool allowResumeAudio then that would be great. I would try to understand why this audio bug happens and then try to make it robust so that pausing or unpausing would not interfere with the audio muting in undesired ways. If we need multiple reasons for muting audio, then consider introducing mute flags that are reserved for different users.

For example MuteAudio(MuteAudioReason_GamePaused), MuteAudio(MuteAudioReason_GameLoading). Something like that.

Note I did not debug this issue so I cannot accurately say what the right course of action now is, but seeing the observations in this pull and the attempted workarounds make it look brittle.

[xezon]

What is the status of this change?

[stephanmeesters]

I will be focusing on this again. Probably will need a separate refactor PR for the audio

[greptile-apps]

Greptile Summary

Replaced direct member variable assignments with proper function calls in GameLogic::reset() to ensure the mouse capture state is correctly unblocked when starting a new game, fixing the camera edge scroll issue after loading a savegame from the in-game menu.

Key changes:

• m_gamePaused = FALSE → pauseGameLogic(FALSE) - ensures proper game pause state reset
• m_pauseInput = FALSE → pauseGameInput(FALSE) - ensures TheMouse->onGamePaused(FALSE) is called to unblock mouse cursor capture
• Changes are mirrored in both Generals and GeneralsMD codebases

Confidence Score: 5/5

• This PR is safe to merge with no risk
• The fix correctly addresses the root cause by ensuring proper cleanup of mouse capture state through the existing pauseGameInput() API instead of bypassing it with direct member variable assignments. The change is minimal, focused, and identical across both codebases (Generals/GeneralsMD).
• No files require special attention

Important Files Changed

Filename	Overview
Generals/Code/GameEngine/Source/GameLogic/System/GameLogic.cpp	Replaced direct member variable assignments with proper pauseGameLogic(FALSE) and pauseGameInput(FALSE) calls to ensure mouse capture state is reset correctly
GeneralsMD/Code/GameEngine/Source/GameLogic/System/GameLogic.cpp	Replaced direct member variable assignments with proper pauseGameLogic(FALSE) and pauseGameInput(FALSE) calls to ensure mouse capture state is reset correctly

_{Last reviewed commit: 1f46bfd}

[greptile-apps]

_{2 files reviewed, 2 comments}

_{Edit Code Review Agent Settings | Greptile}

[greptile-apps]

_{2 files reviewed, 3 comments}

_{Edit Code Review Agent Settings | Greptile}

[stephanmeesters]

I think this should be good to go. The edge scrolling after loading is fixed, and I noticed no audio glitches anymore. Played through some skirmish/campaign/loads and sounds are normal

[xezon]

This means

	m_inputEnabledMemory = TRUE;
	m_mouseVisibleMemory = TRUE;

is set before pauseGameInput is called. Is this correct? Should not be called before?

[stephanmeesters]

I think it's alright, with paused set to false and m_inputEnabledMemory set to true it will trigger an additional call to TheInGameUI->SetInputEnabled(TRUE) however very shortly after it will call the reset function of TheInGameUI and reset any changes anyway

So it should be fine regardless of order but I’ll check if swapping is more efficient here in case m_inputEnabledMemory is often false

[stephanmeesters]

I had another look still think it's OK: input=enabled and mouse=visible are the right default values to enforce, plus I was only ever able to see both m_inputEnabledMemory = TRUE and m_mouseVisibleMemory = TRUE in the debugger when entering reset.