MacOS 14 (Sonoma)

[richmond62]

LC 9.6.3 does NOT run on MacOS 14 beta 4

[richmond62]

So:

2 leading questions to which I would be extremely grateful for 2 direct answers:

1. Will there EVER be functioning builds of OXT for Mac, Win and Lin?

2. Mark Waddingham did 'something' that makes LC 9.6.9 run on the MacOS 14 beta builds: are you aware of what this might
be and are you going to implement something similar?

[micmac]

3. Why has the discussion moved to Facebook when we have this forum

Mic

[richmond62]

Probably because this forum seems like a duck: dead in the water.

While on Facebook one can fool oneself things are a bit more lively.

[richmond62]

Guess what! MacOS 14 sixth beta (oddly labelled as 'beta 4') is no better for the IDE we "know and love."

[richmond62]

Currently my 2018 Mac Mini is "over there" while I use the 2015 27" iMac I got through an appeal, running MacOS 12 like a charm, to hack, piss around with, and generally make a fool of myself with, with the LC 963 IDE . . .

[OpenXTalkPaul]

I wonder about a distribution of OXT that is a Java thing so it can be run on almost any platform?

This is why I am interested in OpenXION and WyldCard and other JAVA based xTalk engines.
I've compiled a 64bit 'native' binary of OpenXION using GraalVM (jdk 17) natifier, but OpenXION is all text based, missing any real GUI, although you can use 'do ... as AppleScript' on Mac, or 'do ... as Python' with Tk for Answer / Ask File Dialog boxes.

Also the Android engine the IDE standalone builder uses is JAVA based (along with being Google Android APIs based), need to take a closer look at that and also the Emscripten Engine. I've done only minor hacking of its JS code, and tested working with it in a Electron app wrapper but I've yet to compile that engine from source (but there's simple instructions available for doing that under Linux ).

[tperry2x]

Judging by the fact that this topic (MacOS 14 Sonoma) has had over 3000 views, it seems to be the most popular topic by far.
If we can work out what is stopping OXT working on Sonoma, that would be a huge positive for a lot of folks it would seem.

I don't have a Sonoma-mac to test it on, but if someone can post the complete output of the crash-dump / diagnostic report, it might have a clue in there as to what the stumbling block is.

[richmond62]

Probably manage that on Saturday.

[tperry2x]

Probably manage that on Saturday.

Hi Richmond. I was just wondering if you had a moment to dig out that crash dump on Sonoma at some point. I can't promise that I'll be able to find the reason it won't run, but I'll gladly go through it and see if anything leaps out at me.

[richmond62]

Oops: sorry: busy day: computer running macOS 14 currently disconnected.

Guess I'll "skip church" tomorrow [God will be pleased].

[tperry2x]

God will be pleased

I'm sure 'God' will forgive you if you can find the required files you need to build your 64-bit mac app in your engine swap.

[richmond62]

I haven't been to Church for about 40 years except for my wedding, my sons' Christenings (which they requested); 15 funerals, and several other people's weddings.

So, one could say that 'God' and I have come to some sort of private arrangement.

HOWEVER, let's push my religious kinks to one side and concentrate on a shared interest.

I WILL connect up my 2108 mac Mini with macOS 14 beta whatever tomorrow and see what's what . . .

I'll get the crash dumps for LC 963 and OXT RC4.

[tperry2x]

Haha, yes - religion and politics. I feel the same way about them as I do the widgets options in the tools palette. (best hidden as an option for those who are interested).

Thanks Richmond. No pressure to get me these, I was just re-reading things I'd said and going through things I might have missed on this forum.

[richmond62]

OK, OK, I am as daft as you thought, I have just tried to Run LC 963 and OXT RC4 on the latest beta of macOS 14 'Soporific', and here they are:

I hope you can understand them: I can't.

[tperry2x]

Thanks for these Richmond.
I've been looking over them, and what jumps out at me (apart from a lot of memory allocation errors), are "MTLCompilerError".
This is probably the MacOS trying to run everything through the metal graphics interpreter, rather than OpenGL.

If I'm correct, I think that changed (was pushed) after version 14 of XCode? If that's the case, it's the Livecode binary looking for dependencies it thinks should be included in the OS.
Of course, these necessary hooks won't be there in Sonoma.

So, the question is - what parts of the engine (or the Livecode binary) rely on the XCode14 dependencies? To know that, I think you'd find these errors if you ran XCode 15 (beta?) on Sonoma and tried to recompile the engine and/or the application binary.

[richmond62]

That probably means we are 'shafted' re getting LC 963 running on MacOS 14 unless Apple does something extremely unlikely before the final release.

[tperry2x]

That probably means we are 'shafted'...

Either that, or if we have a friendly xcode developer who doesn't mind pawing through the compile errors, looking for things as small as signed and unsigned ints? Anyone?

[tperry2x]

This is, of course, all a bit inevitable.
I can't see Apple pulling out all the stops to build in XCode 14 dependencies, to support an architecture they probably consider 'legacy' (or probably even refer to it as 'dead' within Apple internal discussions inside the company). Meanwhile, enterprise and the Windows-way of things (certainly on desktop and server) will continue with Intel/AMD x64 support for a few years. They too will eventually move to ARM I'm sure, due to the lower temperatures and lower energy requirements (a huge plus point in server infrastructure).

This is history repeating. We are on a path to suffer the same fate as Supercard, if we have no native ARM engine support. When Apple does what Apple will definitely do, and drop intel support (currently provided by Rosetta 2), then there will be no way to run XCode 14 and OXT will cease to function on new hardware. After all, they have no incentive to carry on supporting Rosetta and intel if they want to sell ARM hardware to the cool kids - which of course they do.

Even if you and me don't upgrade (and also the many others who can't afford to upgrade), this limits the appeal of OXT for a developer. Why use a tool that can't compile to hardware that other people are potentially already buying?

Meanwhile, Livecode 9.6.8 has native arm support (even if it is deemed 'experimental'), it's something that allows developers to at least release their creations to a wider audience. That's the kicker (as they say). When ARM support is considered the norm, people will regard missing ARM support exactly the way we regard Supercard's missing intel support today.

As in, "what do you mean it doesn't run on an ARM chipset? Why on earth not?"

That might sound a bit bleak, or negative, but I think it's a reasonable conclusion to draw.

To prevent that, we need a developer who is sympathetic to the cause or just passionate about seeing an Open-source version of XTalk carrying on existing. We can still make improvements to everything we need to on intel - we can get everything ready and fine-tuned, so it's as good as it can be on intel. At that point, we are all waiting for an updated ARM engine. Otherwise we will have hit a dead end.

crash-logs.png (240.27 KiB) Viewed 7338 times

[richmond62]

Well, Yes . . .

. . . and it is so typical of Apple to ride roughshod like this without considering anyone's requirements for backward compatibility.