I’d be curious why this is a big project?
You build it already for Windows 64-bit for Intel and ARM64 on MacOS and Linux.
So this sounds like changing whatever scripts/project to build to use the right flags to use ARM64 for Windows. And then fix little things like a missing #if on the go. I know this is some work, but for our plugins, it was a very smooth transition.
And you may it eventually for having a native IDE on Windows for ARM.
When something like this rises up I immediately think in dependencies that escapes their domains, like some 3rd party library not available for some new target, for this case, Windows ARM64. While the author of their dependencies does not provide a solution for such target, they are hands tied or need to write their own code to abandon such dependency.
Now I’m very curious: can you give us a thumbnail sketch of the effort required? I think we sometimes take your team’s effort for granted and it would be illuminating to have some understanding of what has to happen.
Mostly (though not entirely) it has to do with the fact that realtime compilation in LLVM for Windows on ARM is not as mature as it is for other combinations of platform/CPU. Add to this that the marketshare for Windows on ARM is tiny and Microsoft’s interest is waning makes the business case more difficult.
For clarity, if someone came along and was interested in paying us a lot of money to support it, making it make sense from a business point of view, we are fully capable of doing all that is needed which would include additional work on our part for LLVM itself.
Actually Microsoft is putting a whole lot under for Windows Arm those days.
They even ported the whole Visual Studio to it.
I think its very obvious that Microsoft knows that ARM is going to rule the whole mobile and laptop market. MS has done their homework and fixed the Windows for ARM, many major applications and now last the whole Visual Studio.
The missing piece for them is still to let the snapdragon’s exclusive license expire. So they can start selling Windows Arm64. (This exclusive deal dates back to when they attempted the Windows Phone venture, where they partnered with snapdragon at the time). And this is largely the reason why Apple has no bootcamp for Windows Arm64, since Windows Arm64 licenses are not sold (and to buy one you need to buy from partner currently who has snapdragon hardware).
But if googling a bit around then its clear that this exclusiveness contract is thought to expire soon. I think we can expect big things to happen at that time when this market opens up.
Basically this here is the only thing holding Windows on Arm back right now:
They have been but I believe their interest may be waning. When you order a Microsoft Surface 9 for example, the default processor is the Intel core i5. Now that Is the least expensive but the next choice is the i7 which is the most expensive. That’s only then followed by the Q3, their ARM option which is between the two others price wise. You’d think they would be pushing ARM but they don’t appear to be.
Time will tell. If Microsoft is dedicated to ARM and makes more headway with it, then the LLVM team will likely also put more effort into their support for Windows on ARM.