Windows x86_64 -> aarch64 Qt toolchain built on Linux (no Yocto)
-
Hi all --
Long story short, I'm trying to cross-compile Qt for Raspberry Pi. There's myriad articles on this, I'm aware, and I have successfully built Qt 6.4.3 for Linux x86_64 and used that to build a runtime for aarch64 following this article. tl;dr please skip to the last paragraph.
I'm not a fan of Linux desktops, and for a number of reasons I run Windows as my primary machine. I want to be able to develop for my Raspberry Pi using Windows, but build the SDK on Linux. I know this sounds convoluted but I absolutely hate building open-source projects on Windows, largely because of how convoluted it gets.
I was once able to build a Windows x86_64 Qt SDK that cross compiles to aarch64 using Yocto. Once. It took several weeks of troubleshooting and headaches. Not to mention every time I clobbered the build or messed something up it took 24 hours to get back to where I left off.
Can someone please direct me to any resources that may assist me in building a Windows x86_64 -> aarch64 Qt SDK, preferably on a Linux box without Yocto?
Thanks
-
Yes, I'm trying that now, since Windows 11 seems to have better integration than Windows 10 did. Ideally I don't want to be dependent on that, though, especially since I know what I'm after is doable (and I have done it, but I forget how).
This is an ongoing "project" at this point, and I haven't touched it since last year. What I'd like is for someone to be able to articulate how exactly the build works. I understand I need a "host build" initially. From that output I then build the device binaries -- but since Qt is not an "end product" but more of a middleware/development tool, I'm guessing there must be some extra toolchain that it builds as well for when I do actually build an application. So if I were to build on x86_64 Linux -> x86_64 mingw-w64, and then use the output of the mingw-w64 build to generate the device (aarch64 gnu) output...would that work? Or is that not how this operates?
-