Asymmetric encoders, encoder tests.

[tzarc]

Description

Previous asymmetric encoders PR was broken, and was rolled back shortly after merge.
In addition, encoders had tests, but weren't hooked up.

This PR attempts to address both situations -- enabling the existing test runs showed that encoder tests were broken, so I figured it was best to actually fix up encoder sizing assumptions at the same time.

Modifications

• Encoder tests have been re-enabled
• Encoder tests have been added to:
  • Initialisation now clears out the encoder state, fixing commented-out tests
  • Asymmetric tests have been added, with encoder counts: L=R, L<R, L>R, L=0, R=0.
• Encoder implementation has been fixed up such that asymmetric encoders are usable
  • keymap.h has new defines to allow shorthand for encoder counts, as well as per-side
  • Split common has been amended to ensure only the required number of encoders are transmitted
  • Encoder resolutions weren't correctly being set up in keyboards -- we had ENCODER_RESOLUTIONS_RIGHT already documented but there were places where keyboards were using ENCODER_RESOLUTIONS and ENCODER_RESOLUTIONS_RIGHT incorrectly together (or not at all).
  • ENCODER_RESOLUTIONS is used for both left and right if ENCODER_RESOLUTIONS_RIGHT is unspecified, matching ENCODERS_PAD_A.

Types of Changes

• Core
• Bugfix
• New feature
• Enhancement/optimization
• Keyboard (addition or update)
• Keymap/layout/userspace (addition or update)
• Documentation

Checklist

• My code follows the code style of this project: C, Python
• I have read the PR Checklist document and have made the appropriate changes.
• My change requires a change to the documentation.
• I have updated the documentation accordingly.
• I have read the CONTRIBUTING document.
• I have added tests to cover my changes.
• I have tested the changes and verified that they work and don't break anything (as well as I can manage).

[mtei]

FYI:

Last September, I experimented with encoder_read() by rewriting it as follows.

As a result, the estimated execution time of encoder_read() in normal time went from 18 microseconds to 2 microseconds.

bool encoder_read(void) {
    bool changed = false;
    for (uint8_t i = 0; i < NUMBER_OF_ENCODERS; i++) {
        uint8_t new_status = (readPin(encoders_pad_a[i]) << 0) | (readPin(encoders_pad_b[i]) << 1);
        if ((encoder_state[i]&0x3) != new_status) {
            encoder_state[i] <<= 2;
            encoder_state[i] |= new_status;
            changed |= encoder_update(i, encoder_state[i]);
        }
    }
    return changed;
}

[tzarc]

@mtei great, thanks, added to the PR.

[drashna]

FYI:

Last September, I experimented with encoder_read() by rewriting it as follows.

As a result, the estimated execution time of encoder_read() in normal time went from 18 microseconds to 2 microseconds.

...

Can confirm that it doesn't seam to break anything, at least.

Correction. There are some regressions here. It reintroduces the "trigger encoder on startup" issue, again.

[mtei]

Correction. There are some regressions here. It reintroduces the "trigger encoder on startup" issue, again.

You probably have to wait for the input signal to stabilize after the pull-up resistor is enabled.

diff --git a/quantum/encoder.c b/quantum/encoder.c
index 2f37bee6d9..83758a8cbd 100644
--- a/quantum/encoder.c
+++ b/quantum/encoder.c
@@ -121,7 +121,9 @@ void encoder_init(void) {
     for (uint8_t i = 0; i < thisCount; i++) {
         setPinInputHigh(encoders_pad_a[i]);
         setPinInputHigh(encoders_pad_b[i]);
-
+    }
+    wait_ms(1); // After the pull-up resistor is enabled, wait for the input signal to stabilize.
+    for (uint8_t i = 0; i < thisCount; i++) {
         encoder_state[i] = (readPin(encoders_pad_a[i]) << 0) | (readPin(encoders_pad_b[i]) << 1);
     }
 }

EDIT: I posted new PR #16372

[tzarc]

@mtei great, will integrate #16372 into this PR once develop is merged and reopens.

[drashna]

Can confirm that fixes the issue I was seeing with this.