Do the WPILib tools support non-Intel Apple hardware?

[mjhealy]

I'm getting ready to pick up a replacement laptop, but realized that I haven't seen anything to indicate whether the WPILib tools (including the compilers and other components) will run on Apple hardware that isn't Intel-based.

Can anyone help me to determine if M1-based Macs are supported, before I make a "hideous mistake"?

[calcmogul]

They do, kind of. You still need an x86 JDK (WPILib provides one with the installer), and you have to run everything through Rosetta rather than natively.

[mjhealy]

That makes sense for the Java-based tools (e.g., those that are launchable from within VS Code via the WPILib menu), and

I assume that the cross-compiler for C++ (targeting the Rio) is also running under Rosetta (or has actually been ported to M1, though I'm not betting strongly on that), and I'm guessing that we'd need to make sure that Rosetta 2 is installed (e.g., as outlined at the beginning of this post).

Your "kind of" still makes me a little nervous: other than things are executed using Rosetta 2, is there anything specific that needs to be done, or that simply doesn't work as expected? (Now that we're in build season, I'd hate to make the move, and then spend days trouble-shooting problems.)

[calcmogul]

I assume that the cross-compiler for C++ (targeting the Rio) is also running under Rosetta (or has actually been ported to M1, though I'm not betting strongly on that)

To build an M1-native GCC cross compiler targeting the roboRIO, we need a GCC compiler targeting M1. That's still a work-in-progress in GCC 12 (https://github.com/iains/gcc-darwin-arm64), so the only option is running the x86-native cross compiler in Rosetta for now.

Your "kind of" still makes me a little nervous

The tools have M1 support in a "technically works" sense and don't have M1 support in a "troubleshooting support from WPILib" sense. The x86 binaries reportedly Just Work:tm: under Rosetta, but WPILib doesn't test that configuration, so we don't officially provide troubleshooting support if things happen to break in the future.

[mjhealy]

That makes sense, but given that Apple has mostly gotten rid of Intel from their line-up (and is expected to finish the job soon), are there any "best guesses" as to when the cross-compiler port will be available, and support might become more official?

[calcmogul]

That depends on when GCC merges support for M1 targets. It's still a work-in-progress by GCC devs, and they don't have a published timeline, so idk.

The only way to build software for M1 at the moment is using Apple Clang, and the JDK and roboRIO cross compiler build systems both require GCC.

[bobwolff68]

I'll throw in my 2 cents of personal experience here with my M1 MBA for the past year. It works. Both in C++-land and Java-land. I've had zero issues getting the environment installed, updated, compiling, etc. The one area I have not had the opportunity to test is in deploying to the robot itself over wireless or wired connection, but frankly I would be surprised if that were problematic given it's largely network related. --bob - Team 1967

…

On Wed, Jan 12, 2022 at 11:22 AM Tyler Veness ***@***.***> wrote: I assume that the cross-compiler for C++ (targeting the Rio) is also running under Rosetta (or has actually been ported to M1, though I'm not betting strongly on that) To build an M1-native GCC cross compiler targeting the roboRIO, we need a GCC compiler targeting M1. That's still a work-in-progress in GCC 12 ( https://github.com/iains/gcc-darwin-arm64), so the only option is running the x86-native cross compiler in Rosetta for now. Your "kind of" still makes me a little nervous The tools have M1 support in a "technically works" sense and don't have M1 support in a "troubleshooting support from WPILib" sense. The x86 binaries reportedly Just Work™️ under Rosetta, but WPILib doesn't test that configuration, so we don't officially provide troubleshooting support if things happen to break in the future. — Reply to this email directly, view it on GitHub <#3902 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAFYP5TUP2WAXWF2BM24K5DUVXIJBANCNFSM5LZOEANA> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.*** com>

[mjhealy]

Thanks, Bob: that goes a long way towards establishing a comfort factor for me. :-)

[spencer-gre]

Deploying code works fine with M1 MB. Only thing you won't be able to do is run the robot using the official driver station, although I heard there are unofficial cross-platform driver stations that do work. Have not tested them though.