#964 relates to AAX, but the comment in IGraphics_view.mm mentions AUv2 related code. Have you got a reproducible bug for this that doesn't involve EUCON/AAX?

It isn't very nice. Something about it doesn't feel quite right.

Could we not have an IGraphics::mIsClosing variable and test it, via a getter in IGEditorDelegate::CloseWindow() ?

I've made the requested changes.

The comment about AUv2 is yours previously, and not mine - however, this part of the call structure interacts to create the issue. Essentially in AUv2 this is how we find out that the user has closed the window and thus destroy the IGraphics.

The EUCON error relates to two consecutive calls to the delegate OpenWindow() and the crash can be reproduced just by calling it twice in a row in the app host.

In this scenario the following happens:

• OpenWindow() in IGraphics calls CloseWindow() in IGraphics to free old resources
• CloseWindow() in IGraphics removes the view
• That calls back into the delegate to CloseWindow() destroying IGraphics
• We are still in the OpenWindow() call of the IGraphics

We could do as you say, but this is a Mac only issue and:
1 - it seems weird to propagate that up to the delegate
2 - it isn't that we are closing in general - we are closing at the IGraphics level (keeping IGraphics open) so putting it there is potentially confusing.
3 - there's already a closing check in the delegate

Another possibility is that we adapt the delegate to ignore calls to OpenWindow() if the window is already open (removing the CloseWindow() calls within OpenWindows() IGraphicsMac and IGraphicsIOS) - is that a better pattern?

Another possibility is that we adapt the delegate to ignore calls to OpenWindow() if the window is already open (removing the CloseWindow() calls within OpenWindows() IGraphicsMac and IGraphicsIOS) - is that a better pattern?

That does sound a bit nicer. I think i'd prefer it!