Create an xtask subcommand to generate eFuse fields

[jessebraham]

Rather than keeping the (now outdated) CSV files in this repository, we can instead just generate them periodically using the latest versions from ESP-IDF. There's no need to parse these files in the build script during every build.

Since our CSV files were out of date, I ended up pretty much completely rewriting the parser. I think it's easier to understand and is a bit more flexible now, so probably a win anyway.

There is still room for improvements here, we're still skipping a few fields, but I'm not sure how relevant they really are in practice.

Did some small refactoring to accommodate this, nothing noteworthy though.

[bjoernQ]

I think it's a good thing to not regenerate the efuse fields during the build - they are not expected to change.

The idea is to use the CSV from esp-idf directly when generating, right? I agree there is little use in copying them

I remember the old parser had some problems with the MAC address for P4 which needed some modification of the original CSV but seems your changes fixed it as far as I can see - generally it's better to have a parser which doesn't need random modifications to esp-idf's CSVs to work 😄

[JurajSadel]

LGTM, thank you for doing this!

[jessebraham]

I remember the old parser had some problems with the MAC address for P4 which needed some modification of the original CSV but seems your changes fixed it as far as I can see - generally it's better to have a parser which doesn't need random modifications to esp-idf's CSVs to work 😄

Yes this was also a problem for the ESP32 when trying to generate from the latest eFuse tables, which is what ultimately led me to rewriting the parser. This is handled by a second processing iteration over the fields once they've been parsed.

[MabezDev]

LGTM, thank you!