applet.sensor.qmc5883p: initial implementation

[i-infra]

Initial implementation for the qmc5883p magnetometer sensor - i2c @ 0x2C.

This code is tested working with the GY-271 breakout ie) https://www.aliexpress.us/item/2251832421287528.html.

Shoutout @adafruit (referenced their libs)

Datasheet for part @ https://www.qstcorp.com/upload/pdf/202512/2C939E5AA0704285BC3BE71132B8629B.pdf

PS: sending sm love to the glasgow team <3

[whitequark]

At a glance this looks great! The only issues I can see without closely looking through the functionality is ones that were already a problem in the codebase before you started working on your applet.

Please do the following in addition to inline comments:

• Add yourself to CODEOWNERS for this applet
• Add a documentation page at least for the CLI
• Add types to the Interface class (if this is something you would normally use in your code)
• Document the Interface class and add an API documentation section to the documentation page (I will merge the applet without this but it is preferred to have API docs for new applets like this one)

[whitequark]

Regarding code formatting, there is a Ruff configuration and a pre-commit file that invokes it, this configuration is the one appropriate for Glasgow source code.

[whitequark]

I did a rough pass (not a final one), looks better. By the way, I'm not sure why but "requesting review" from me on GitHub doesn't seem like it sends me an email; leaving a comment is the best way to get my attention.

[i-infra]

I think it's good to go? Need to figure out legacy vs appletv2 stuff eventually, but check it out when you get a sec! Suspect it's approaching mergable.

[whitequark]

Need to figure out legacy vs appletv2 stuff eventually, but check it out when you get a sec! Suspect it's approaching mergable.

Oh I somehow missed that you were still using the legacy applet infrastructure. New legacy applets are a big "no", I've been trying to migrate all of them to the new infrastructure for months already.

[i-infra]

@whitequark I think this is good for re-review!

[whitequark]

Thanks, will take a look once I'm done with some urgent bureaucracy.

[whitequark]

Could you also use an enum here?

Also, generally I avoid making things such as "register enumerations" IntEnums unless there's a specific reason to do so; accidentally doing something like _XOUT_LSB + 1 (in a more complex code path) and getting a nonsense result hurts a lot when debugging.

[whitequark]

I think you're missing the enum here, unless it's intentional.

class Register(enum.Enum):
    CHIPID = 0x00
    XOUT_LSB = 0x01
    ...

[whitequark]

This should be mergeable after the last review is addressed.

[whitequark]

Ah sorry, missed two really important points I stated elsewhere:

• needs documentation (you should include both the CLI docs as well as the Interface class and every enumeration it references; check that there aren't dead links in the generated docs)
• needs tests (at a minimum an assertBuilds but I strongly prefer having record-replay tests for at least the basic functionality; should be easy here

[whitequark]

Should be mergeable after the latest round of changs and maybe some tests!

[whitequark]

I think you're missing a lot of the newly added enumerations (the ones that now get surfaced in the public API).

[whitequark]

I think you're missing the enum here, unless it's intentional.

class Register(enum.Enum):
    CHIPID = 0x00
    XOUT_LSB = 0x01
    ...