Struct still crashes in some cases

[EthernautParticipant]

This problem on Github is marked as closed, but it still crashes when the struct contains an Image field.
https://github.com/blitz-research/monkey2/issues/114

Monkey

#Import "<std>" #Import "<mojo>" Using std.. Using mojo.. Function Main() Local testStack := New TestStack For Local n := 1 To 20 Local newItem := New TestItem testStack.Add( newItem, Rnd( 0, 100 ) ) Next For Local n := Eachin testStack.items Print( n.depth ) Next End Class TestStack Field items := New Stack< TestItem > Method Add( item:TestItem, depth:Float ) item.depth = depth items.Push( item ) End End Struct TestItem 'Change this to a Class and it will compile! Field depth:Int Field img: Image 'Comment out this line and it will compile! Field color: Color End

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

#Import "<std>" #Import "<mojo>" Using std.. Using mojo.. Function Main() Local testStack := New TestStack For Local n := 1 To 20 Local newItem := New TestItem testStack.Add( newItem, Rnd( 0, 100 ) ) Next For Local n := Eachin testStack.items Print( n.depth ) Next End     Class TestStack Field items := New Stack< TestItem > Method Add( item:TestItem, depth:Float ) item.depth = depth items.Push( item ) End End     Struct TestItem 'Change this to a Class and it will compile! Field depth:Int Field img: Image 'Comment out this line and it will compile! Field color: Color End

[abakoboParticipant]

Images are initialised by NewAppInstance.
See this closed issue: https://github.com/blitz-research/monkey2/issues/103

Edit: opps it’working with class so it must be something else?

[EthernautParticipant]

Yeah, this doesn’t even compile, so I think it’s something else. There’s no App.Run in this minimal example, but when I first ran into it I had a proper AppInstance.

[Mark SiblyKeymaster]

Fixed now in develop branch. I forget that assigning a struct actually also involves a bunch of member assignments.

This is a hard one to get 100% right though, as there’s a tension (in c++) between not wanting to #include too much, which can cause forward referencing issues, and having to #include enough to keep compiler happy. And fixing one always affects the other…

I think a new approach may be in order and have some ideas but for now I’ll just keep dealing with these on a case by case basis.

[EthernautParticipant]

<edit>
Never mind, works fine now. Thanks Mark!!!

[Author]

Posts