On the one hand it is good to discover that someone is tackling getting TianoCore working on the Raspberry Pi 5.
On the other hand, they still have the destructive backspace behaviour, and inefficient recursive implementation, that breaks the boot loader spinners that the NetBSD and other boot loaders display. It's a tiny thing, but if one is used to the boot sequence the absence of a spinner makes the experience ever so slightly jarring.
I should add, by the way, that this nicely demonstrates M. Geerling's point here about catching bugs by running things on a Pi.
The TianoCore's unnecessarily recursive implementation of a destructive BS is slow enough, on a Pi 4, and in combination with how the boot loaders themselves emitted their spinners, that I could just, very occasionally, see parts of spinner characters flashing very briefly on the screen when the frame refresh timing was just right; which led me to look into what was going on.
Wasn't there a post somewhere on HN yesterday about how slowing down your programs can help you catch problems? Using low end hardware is an automatic way of forcing that :)
On the other hand, they still have the destructive backspace behaviour, and inefficient recursive implementation, that breaks the boot loader spinners that the NetBSD and other boot loaders display. It's a tiny thing, but if one is used to the boot sequence the absence of a spinner makes the experience ever so slightly jarring.
* https://github.com/NumberOneGit/edk2/blob/master/MdeModulePk...
* https://github.com/tianocore/edk2/blob/master/MdeModulePkg/U...
* https://tty0.social/@JdeBP/114658278210981731
* https://tty0.social/@JdeBP/114659884938990579