[WIP] Ubuntu 24.04 for MinGW

[FyiurAmron]

This has been separated from the mingw-std-threads removal PR, since I found out it's possible to easily switch to POSIX on 20.04 (and earlier as well I think).

related tickets/blockers:

done:

• Partial sync to upstream calf-studio-gear/calf veal#13 = overloaded virtual methods in plugins/LadspaEffect/calf/veal/src/analyzer.cpp etc. break compilation on newer g++ #7284 & update veal submodule #7290
• Fix to allow NSIS packaging to work for non-MSVC Windows builds #7287
• Fix for Werror=self-move in test for GCC >= 13 #7288
• libgig future vs error=overloaded-virtual and other problems #7291 & Suppress warnings in third-party code #7319
• minor cleanup of MinGW-related scripts #7327 as a stepping stone to make pinpointing the 32-bit issues easier

pending:

• identify the source of 32-bit cross-compilation problem (TBH seems like old lib versions clashing with new ones, STK/FLTK and Qt probably)
• use Use vcpkg for Linux CI #7316 (or actually its MinGW counterpart)

[PhysSong]

FYI, There is https://github.com/LMMS/lmms-ci-docker for container images

[FyiurAmron]

@PhysSong yes, as you can see in both the previous and current commits I've been tweaking the code from there directly (current content is an edited copy-paste of what we have there). I do however recall people talking about ditching the CI repo, and anyhow I see no sense in going through the same code in two separate repos for a draft, if I can just do the tryouts here directly. ATM the only problem is with compatibility with 32-bit MinGW with 24.04 (mostly due to our outdated version of libfltk if I get it correctly from the logs), and that's the only thing blocking the upgrade right now, AFAIK. As soon as I have the code working here, I'll submit a matching PR to lmms-ci-docker repo, if that will be the solution-of-choice at the time I'm done :D