Download the artifacts for this pull request here:

GUI:

• GUI-windows-latest-isBundled-false-isSingleFile-false
• GUI-windows-latest-isBundled-false-isSingleFile-true

CLI:

• CLI-macOS-latest-isBundled-false
• CLI-ubuntu-latest-isBundled-false
• CLI-windows-latest-isBundled-false

Note that there's currently a bug, where if you leave null entries exposed on the asset tree for too long and move around too much, maybe click into one of the nulls (the true conditions are uncertain), it could cause a hang or even crash. I'm consistently getting the crash as descripted by the incomplete stacktrace below:

System.ArgumentNullException: Value cannot be null. (Parameter 'key')
   at System.ArgumentNullException.Throw(String paramName)
   at System.Collections.Hashtable.ContainsKey(Object key)
   at MS.Internal.WeakDictionary`2.TryGetValue(TKey key, TValue& value)
   at MS.Internal.Helper.SetItemValuesOnContainer(DependencyObject owner, DependencyObject container, Object item)
   at System.Windows.Controls.VirtualizingStackPanel.InsertContainer(Int32 childIndex, UIElement container, Boolean isRecycled)

Might be a WPF bug, if inquired I will try to grab a full crash dump.

Note that there's currently a bug [...] (the true conditions are uncertain) [...]

Ok now I see, you have to very specifically click into the null entries first

What's even funnier is, if I run ModTool through VSCode, it hangs instead of crashing

Sorry to bother you, but it appears there are two issues:

1. Shaders are improperly imported creating null entries that cause shaders not to be visible in editor and create errors when saving code (Seen in the Mindwave Demo)
2. Structs and/or variable set to functions are improperly handled, and not being properly recompiled. This can been seen by trying to import modified minigames from Mindwave Demo into their data.win.

oh thank goodness

Code does not work for me, Decompiling just gives me this. /* EXCEPTION!
Underanalyzer.Decompiler.DecompilerException: Decompiler error during AST building: Data left over on VM stack at end of fragment.

Code does not work for me, Decompiling just gives me this. /* EXCEPTION! Underanalyzer.Decompiler.DecompilerException: Decompiler error during AST building: Data left over on VM stack at end of fragment.

@MichaelEpicA Unsure if this is fixed in upstream Underanalyzer (this PR may need to be updated), but does this happen on ALL code entries, or just some? If it's just some, if you could send a small code entry's disassembly, that would be helpful for debugging (but report it over here: https://github.com/UnderminersTeam/Underanalyzer/issues). Also, you may be able to turn that specific error into a warning in the UTMT settings.

Can confirm this can be turned into a warning in the UTMT settings, and saving appears to work, however I'm doubtful it will load correctly due to UMT reporting a object miscount. It is every single code entry.

Can confirm this can be turned into a warning in the UTMT settings, and saving appears to work, however I'm doubtful it will load correctly due to UMT reporting a object miscount. It is every single code entry.

@MichaelEpicA Hmm... if there's a small code entry sample, I guess I'm still curious to look at its disassembly to see if it has to do with actual decompilation, or if there's some other problem at play here.

If so, should I just move this over to https://github.com/UnderminersTeam/Underanalyzer/issues then?

If so, should I just move this over to https://github.com/UnderminersTeam/Underanalyzer/issues then?

Sure, I'll probably just close it if it ends up being irrelevant.

Note: the editor (on Mindwave's Demo) shows shaders now
However, when AddAssetsFromList(Data.Shaders, RefType.Shader); is called, it still crashes when compiling.
Also, when attempting to compile code like this:
define_boop("boop", { init: init, fleas: fleas })
It causes an error at 1499 of AssemblyWriter.CS.

While the editor can now save the define_boop("boop", { init: init, fleas: fleas }), it appears there are other issues that appear upon runtime after recompilation including SetStatic not being readded and global scopes being changed to static scopes. This is visible in the Mindwave Demo when attempting to recompile a microgame.
Recompiled is on the left, Original is on the right

Error:

___________________________________________
############################################################################################
ERROR in action number 1
of  Step Event0 for object transition_controller:
Variable <unknown_object>.ms0scr_player_tick(103885, -2147483648) not set before reading it.
at gml_Script_mg_sGrizz
############################################################################################
gml_Script_mg_sGrizz (line -1)
gml_Script_anon@6949@gml_Object_obj_minigame_controller_Create_0 (line -1)
gml_Script_anon@5610@charselect_character_constructor@charselect_character_constructor
gml_Script_anon@4882@anon@4418@charselect_character_constructor@charselect_character_constructor
gml_Object_transition_controller_Step_0

While the editor can now save the define_boop("boop", { init: init, fleas: fleas }), it appears there are other issues that appear upon runtime after recompilation [...]

For now I'm waiting for the new UA-provided compiler to be integrated into the upstream underanalyzer branch, which may or may not solve this done

NTS: Two things to check & fix after I come back to this

• Bogus "if asset list is null" check removals done
• Whether some hacks in CheckForGMxxxx functions I added here are necessary
  • Update: I think the SOND 2024.6 check mod is necessary because it's not realistic to detect 2024.8 in a YYC game