Update tock-registers dependency to 0.8

[berkus]

Update for the new version of tock-registers.

[andre-richter]

Lazy question from my side: Will it break anything?

[berkus]

Does not seem so, the only breaking change touches try_from -> try_from_value and that is not used in cortex-a.

Can't really test on my OS, because cortex-a dependency on 0.7 breaks the build so far (too lazy to do manual overrides, don't ask).

[andre-richter]

I am currently pondering if this justifies a new major release of cortex-a, since the user is encouraged to bump the tock-registers dependency of the user's project in tandem, and it contains a breaking change.

[berkus]

I believe it does, semver and all that...

[andre-richter]

I am seeing some issues with the 0.8 tock version when trying this out locally (tock/tock#3228). Please give me some time to dig around before we indirectly force people to switch to it due to cortex-a.

[berkus]

@andre-richter alright, I will get to testing it more with my code this week as well.

[andre-richter]

Let’s go directly to 0.8.1, so pending tock/tock#3242.

Since it will be a breaking change due to the tock upgrade anyways, I’ll take the opportunity to also include 9421f40. This has been bugging me for some time now.

[berkus]

👍🏼

[berkus]

Well, my code here broke because of tock/tock#3088 - making this a major release was a good call.

Error is

     |           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ the evaluated program panicked at 'Invalid size for struct RegisterBlock (expected 244, actual struct size differs)', machine/src/platform/rpi3/gpio.rs:33:1
     |
    ::: /Users/berkus/.cargo/registry/src/6github.com-1ecc6299db9ec823/tock-registers-0.8.1/src/macros.rs:165:1

So it's conveniently reports what I already know (the address of @EnD marker, but doesn't tell what it should be, in its opinion). Should be easy to fix.