here is what caused that issue:
when you save the file for the very first time, your identifier is exactly the way you type it in in the text field.
here's mine:
com.BkTools.blockingDoc
when you save changes to a file, the identifier is lowercased:
com.bktools.blockingdoc
I actually test for my type identifier, because I will probably need to have more than one.
So I uses a Switch. and the list of options needs to be exhaustive, so I have a Default: option that returns an empty Data() But that should never get called. under any circumstances. I have a closed loop here. I am saving MY file type, I Am opening MY file type. I have written down MY identifier, so I can Expect MY file identifier.
except. I can't. Somebody transforms my identifier before I get it. which is a cardinal sin.
WORSE,
they transform my identifier, only SOME of the time.
which in my book is far far worse a cardinal sin.
the workaround... when I inevitably run into this problem again and there's no documentation then either:
is to either use all lowercase in the identifier, or to use the convenience method : lowercase() on the string in the switch and just check for the lower case version.
half a day looking for that little gem.
Post
Replies
Boosts
Views
Activity
that is the error message generated when it crashes. the AMD Radeon Pro W5700X 16 GB is apparently incapable of raytracing.
the net result was:
the book I was following made assumptions that turned out to be wrong. The code in the book could never work.
the Matrix4x4 structure in the book assumes a row layout in the the float4x4 used in Metal. It doesn't work like that. it uses a Column layout.
ie: the matrix I built from the tutorial assumed that each simd_float4 passed in was going to be a row. instead it was a column.
this screwed up everything.
but I also had issues with the matrix weirdly breaking my identity values.
that screwed up everything.
and my rotation function, was faulty.
thanks Quinn, I "think" I was worried over nothing. As I recall, I was worried that In the plugin itself, I assumed that getting the "main" bundle I'd wind up with the Plugin's bundle. But, I tried it. and it seems to give me the Host Application's Bundle. Which is what I wanted, but did not expect.
var.if var envelope = theParent as? CNEnvelopment{
there is a parent class to the classes that are CNEnvelopment.so the child classes are something like :public class CNNode : CNDagObj, CNIdentity, CNProperties, CNEnvelopment, CNStackObj, CNDagConnection, BKObserver, CNRenderable, CNOutputsObj {CNDagObj is the parent class. Not all CNDagObjects are CNEnvelopment.the Property in which we find the CNEnvelopment, is of the class CNDag.public var parent : CNDagObj? = nilall objects that will be a parent, will also be CNEnvelopment, But as I recall... the compiler doesn't like it if I do this instead :public var parent : CNEnvelopment? = nilso I'm stuck with the next lowest common denominator class (CNDag) i suppose i should :if let envelope = theParent as? CNEnvelopment{?
not for me. the app actually compiles and runs in the OS. This is a BUG in XCode.
having the same issue. it ain't me.
Thanks OOPer,i knew it was going to be easier than what I was doing.youre a lifesaver.
turns out: no. it doesn't work. not at all. SteamVR mac seems to have been abandonded in Beta, and it's the only option for VR on the mac. the headset seems to work just fine. aspects of SteamVR seem to work just fine... anything that presents content, is thoroughly buggy. and it's buggy in a way that makes me think it was sabotaged.
Any word on this situation getting better?vive pro here with a 2019 mac Pro. the controllers firmware updated after numerous tries. Nothing else did.having really unacceptable problems with the head unit, but I cannot even perform the first step in troubleshooting: updating the firmware.