Improve loclass logic for readers doing keyrolling.

[nvx]

Improve loclass logic for readers doing keyrolling. Writing to the loclass log is also now deferred until the CSN changes for speed.

What's new

Removes the forced anticoll between MAC collections except for when the CSN changes. This speeds up the collection process and should be more reliable with readers doing keyrolling.

Verification

• Perform loclass and ensure it is still able to recover keys from a legacy elite keyed reader.

Checklist (For Reviewer)

• PR has description of feature/bug or link to Confluence/Jira task
• Description contains actions to verify feature/bugfix
• I've built this code, uploaded it to the device and verified feature/bugfix

[bettse]

I don't claim to know all the details of loclass, but overall the codes makes sense.

[bettse]

I used the nfc-iclass project to make a "KDE" keyroll config card, which I used on a Rev A reader. I ran the new code against it, downloaded the file, and was able to use https://loclass.ericbetts.dev/ to calculate the key I'd configured 😁.

[nvx]

Fixed that bug with the numbering in the loclass log being wrong and also made the code a little more defensive in the emulation loop to stop when finished rather than relying on the UI to stop the emulation.

[bettse]

Pulled down the new code and used it again with my keyrolled reader. So fast! Log looks good and works with loclass.ericbetts.dev. I'm going to use my new found poweres to approve try the "Approve and run" in the PR.

[bettse]

Looks like I can approve, but merging is blocked 👍

[skotopes]

@bettse looks like you are missing in codeowners. Can you make PR with it?

[skotopes]

Also don't forget changelog + catalog update

[bettse]

We've also got #47 open, so we could do the bump after that

[skotopes]

okok