fix(headless): Skip early load abort in headless replay simulations

[githubawn]

Fixes the CI replay checker instantly completing introduced by #2336.

Bypasses the QuitGameException in GameLogic::updateLoadProgress when running in headless mode. This ensures that any early initialization warnings that set the quitting flag do not prematurely abort the replay simulation during map loading.

Claude helped. Worked locally. Let's see what it does on our CI.

[greptile-apps]

Greptile Summary

This PR fixes a CI regression introduced in #2336 where headless replay simulations would immediately abort due to WM_CLOSE firing before TheMessageStream was ready, inadvertently setting the quitting flag and causing GameLogic::updateLoadProgress to throw QuitGameException. Changes are applied symmetrically to both the Generals/ and GeneralsMD/ trees.

• WinMain.cpp (both trees): The WM_CLOSE fallback path that calls setQuitting(TRUE) is gated behind !TheGlobalData->m_headless, so a spurious close message during early engine init no longer poisons the quitting flag for headless runs.
• GameLogic.cpp (both trees): updateLoadProgress adds the same headless guard as a second line of defence, ensuring that even if the quitting flag is set through any path during map loading, a QuitGameException is not thrown while running headless.

Confidence Score: 5/5

Safe to merge — changes are narrow, well-commented, and symmetric across both game trees.

The fix correctly gates two independent code paths (WndProc and updateLoadProgress) behind the existing m_headless flag. The !TheGlobalData null-guard is the right conservative fallback: when global data is unavailable the behaviour reverts to the original (throw/setQuitting), so non-headless production builds are unaffected. No new state is introduced and the change does not touch any shared data structures.

No files require special attention.

Important Files Changed

Filename	Overview
Generals/Code/Main/WinMain.cpp	WM_CLOSE handler skips setQuitting(TRUE) in headless mode when TheMessageStream is not ready, preventing early quit-flag poisoning during engine init.
Generals/Code/GameEngine/Source/GameLogic/System/GameLogic.cpp	updateLoadProgress now skips QuitGameException in headless mode, acting as a second-line defence against spurious quit flags during map loading.
GeneralsMD/Code/Main/WinMain.cpp	Identical fix to Generals/Code/Main/WinMain.cpp applied to the Zero Hour tree; no functional differences.
GeneralsMD/Code/GameEngine/Source/GameLogic/System/GameLogic.cpp	Identical fix to Generals GameLogic.cpp applied to the Zero Hour tree; no functional differences.

Sequence Diagram

sequenceDiagram
    participant OS as Windows OS
    participant WP as WndProc (WinMain)
    participant GE as GameEngine
    participant GL as GameLogic::updateLoadProgress
    participant CI as CI Replay Checker

    Note over OS,CI: Headless replay startup — engine still initialising

    OS->>WP: WM_CLOSE (early, TheMessageStream not ready)
    alt Non-headless (old behaviour)
        WP->>GE: setQuitting(TRUE)
        GE-->>GL: "getQuitting() == true"
        GL->>CI: throw QuitGameException() — simulation aborted
    else Headless mode (new behaviour)
        WP-->>WP: skip setQuitting (headless guard)
        Note over GL: updateLoadProgress also guards: skip QuitGameException if headless
        GL->>CI: loading continues normally
    end

    Note over CI: Map loading completes, replay runs to end
    CI->>CI: Simulation finishes, process exits cleanly

Loading

_{Reviews (5): Last reviewed commit: "another try" | Re-trigger Greptile}

[githubawn]

did some triaging and it happens much earlier than found, run 5930 instead of 7153 https://github.com/TheSuperHackers/GeneralsGameCode/actions/runs/22267506812/job/64416452283 basically either the first or second commit of the PR

[xezon]

As long as the test passes green, it is not working correctly.

[stephanmeesters]

Suggest to first revert #2336 entirely, observe tests (of GitHub CI), then divide and conquer