pippin authored 1 year ago and q3k committed 1 year ago

There seems to be a leak somewhere, causing trouble if you have
more than 16ish apps, this is merely a stopgap for initial
power users.

This was useful during development, but is just a nuisance when trying
to use the badge REPL now.

schneider authored 1 year ago and dx committed 1 year ago

No automatic switching of anything, just expose it to MP.

Closes #16

Also:

- improved LED docs
- fixed the repr of the leds module
- fixed fix LED_POSITIONS in the sim

About the last point, the previous version was clockwise but not
starting from the top, but from the "first" LED according to the
hardware, which is ~11 LEDs after the one at the top. The C firmware
already adjusts for this.

Not fixed by this commit: Melodic demo (it's already a mess)

Didn't need changes: nick (now rotates clockwise)

q3k authored 1 year ago and schneider committed 1 year ago

The menu/application button naming never gained traction and I don't
think it will.

Let's just call them 'left' and 'right' but allow the user to swap them
in settings.

This prevents the state of initial color/text-align/line-width/etc to
be dependent on the state of the previous rendering call - presuming
further nested balanced save/restore pairs.

Introduces an update function to the BSP IMU implementation which
fetches the complete state of the BMI270 at once. This includes memory
range of the auxilary pressure sensor.

Introduces a task in st3m which updates the sensor readings every 10 ms.
Priority is chosen to be below captouch and audio.

BSP functions let the user know if there actually was a new reading on
the specific sensor. st3m always return the last measurement without
this extra information.

Slight adjustment of gradient stops to distribute primary colors better
on characters of flow3r. Use single precision constants rather than
double precision constants for coordinates.

1. We weren't setting bProtocol to IAD (technically required for CDC-ACM
   devices which have IAD)
2. We weren't counting the IAD pseudo-interface as one of the two
   interfaces in the configuration.

This reverts commit 4f0a4b80b4137e1a192111fd330eb819d45c8764.

This is a first-pass implementation of an SD card driver, implementing
both low-level block-based access and mounting to VFS.

Short reads generally shouldn't happen, but let's be defensive there.

Instead, let's explicitly check the return code of i2s reads.

Brought to you by: “my interrupt didn't get initialized and i got very
confused”.

This moves the raw read/write access code from the recovery mode into
st3m, and gives us the ability to mount/unmount the FAT partition on
demand.

Tested by doing the following in REPL:

    >>> import hardware
    >>> ctx = hardware.get_ctx()
    >>> ctx.rgb(1, 0, 0)
    <ctx_type>
    >>> ctx.rectangle(-120, -120, 240, 240)
    <ctx_type>
    >>> ctx.fill()
    <ctx_type>
    >>> hardware.display_update(ctx)

Fixes #25.

Whoops.

This allows us to os.listdir("/").