What if they fix errata?

[SpenceKonde]

In theory, he errata wil some day get fixed, If we are luck, this will occur before our sun expands into a red giant in severak billion years incinerating all life on earth. This would make sourcing fixed parts difficult, as well as raisin new cooling challenges, not to mention fact humanity would be burnt to a crisp, which would likely adversely impact the availability of support n the event that the errata were not entirely fixed.

We can only determine the REVID for sure at runtime.

But uh.. like it has huge consequence. For example, on AVR DA/DB, how do we support mux-aware pwm?! The digital_pin_to_timer table has to be completely different, bt that can't be changed a runtime, only at compile time when we don;t know the information we need.

Or we rely on high overhead hackjobs that constantly check revid (but storign it in a vairiable is no better. - SRAM or SFR, LDS still three damned clocks.

Or we steal one of the GPIORs, so at least the tests take 1 clock cycle instead of 4....

On the other hand, most of the errata right now are literally non-issiues

One other approach is to have tools submenu where a user must specify it.
I wonder if SerialUPDI could be made to detect it and refuse to upload to a revid that we could tell was incompatible with what is about to be uploaded

How do folks think we should handle tis if the time ever comes? The double-D's themselves will have no sch problems. They can be differentiated by virtue of part number, and it wasn;t a problem with the A4 DB's because so few made it out the door. But this will be a problem for DA and DB (also tinyAVR, though those bugs are mostly of the "This won't work unless you have fixed silicon, sorry" variety, whereas on Dx there are a whole bunch of places where issues get worked around...