Subsurface on Apple Silicone

[Dirk Hohndel]

> On Oct 15, 2021, at 9:55 AM, Robert Helling wrote:
> 
> Hi,
> 
> has anybody managed to build Subsurface on Apple Silicone? Because I haven’t.

Yeah, I still don't have a Mac with M1 silicon, so no, haven't even tried... my understanding is that the x86 version works just fine, though...

> Here is a log of my frustrations.
> 
> Before leaving for the summer vacation, I received a new MacMini at the office featuring the new ARM based Apple hardware that everybody is so excited about. It runs Big Sur.
> 
> It took several weeks to find the time to properly take care of it, install all the software I need for work, get permissions right (it seems they changed what admin users can do when the machine is part of an active domain or better what they cannot do, for example install software updates etc etc). But those things I think I have under control by now.
> 
> So I move forward and tried to set up a build environment for Subsurface. I encountered halve installed versions of Xcode which on the other hand leave behind zombi installations when removed that confuse auto build when still around when there is a second version as well (this took me two days to understand). But also that is solved. I think.

This might be good to document in more detail, especially how to fix it (for the next poor person coming around to this)


> But now I hit a road block: I am able to build the subsurface object files. But then I came to the understanding that Qt 5.x does not exist in version for Apple’s ARM processors, only for the 86 family (which could be used in Rosetta mode and indeed the official Mac binaries from Dirk’s server work fine on that machine). But I have not been able to convince the Subsurface build system to build 86 family binaries on my ARM machine.

No idea how to do that, frankly. I'm sure you've found this page https://www.qt.io/blog/qt-on-apple-silicon <https://www.qt.io/blog/qt-on-apple-silicon> and if you wanted to experiment with Qt6 there's more stuff here https://doc.qt.io/qt-6/macos.html <https://doc.qt.io/qt-6/macos.html>

I think eventually we will need to port to Qt 6 - and that should fix the core of your problem, right?

> So I see three options to move forward: „Cross-build“ Subsurface for the old hardware platform. Or wait for ARM built binaries of Qt 5.x (which as I understand might never happen at least in the open source world). Or move Subsurface forward to work with Qt 6 (which might be in better shape w.r.t. ARM but I don’t really understand the current situation in detail).

I think you have that right. Allegedly one can build "fairly decent" Qt 5.15 binaries for M1. Just takes a while :)

But really, the only logical path is Qt6

/D

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.subsurface-divelog.org/pipermail/subsurface/attachments/20211015/05fbbcdb/attachment.htm>