Adafruit 'seesaw' support #612
Conversation
seesaw/seesaw_test.go
Outdated
| assertDeepEquals(t, err, nil) | ||
| } | ||
|
|
||
| func assertDeepEquals[e any](t *testing.T, actual, expected e) { |
There was a problem hiding this comment.
I tried quicktest again, does not compute for me. I.e. even the existing usages on dev don't work locally.
There was a problem hiding this comment.
even the existing usages on dev don't work locally.
That seems to imply something wrong with your local config. no?
There was a problem hiding this comment.
Very fair point :)
- I now went through re-building tinygo from v0.30.0 and still get the same issue:
~> tinygo test -v ./tester
panic: unimplemented: (reflect.Type).IsVariadic()
FAIL tinygo.org/x/drivers/tester 1.314s
~> tinygo version
tinygo version 0.30.0 linux/amd64 (using go version go1.20.8 and LLVM version 16.0.1)
- then I had a look at the workflow on the repo that runs it successfully, if I do the same with the docker image I still get the error:
> docker run --rm -it -v /<redacted>/tinygo_drivers:/build ghcr.io/tinygo-org/tinygo-dev:latest /bin/bash
root@2632e6b72f70:/go# cd /build
root@2632e6b72f70:/build# tinygo test ./tester
go: downloading github.com/frankban/quicktest v1.10.2
...
go: downloading golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543
panic: unimplemented: (reflect.Type).IsVariadic()
FAIL tinygo.org/x/drivers/tester 1.366s
- furthermore I found that even the GitHub Action on the
devbranch seems to succeed with failing tests:
I still don't know what's wrong with my environment, but I found a fix for the CI action not properly failing: #614
Any pointers? At least the docker environment should work, no?
There was a problem hiding this comment.
Ok, I figured it out 🤦
Apparently the way to run unit tests is to use go test and not tinygo test.
Out of curiosity, what's the purpose of tinygo test then?
It runs tests fine without the quicktest dependency :D
trichner
force-pushed
the
seesaw2
branch
2 times, most recently
from
7327a55
to
badc0de
Compare
seesaw/seesaw.go
Outdated
|
|
||
| // This is needed for the client seesaw device to flush its RX buffer and process the command. | ||
| // See seesaw datasheet for timings for specific modules. | ||
| time.Sleep(delay) |
There was a problem hiding this comment.
Was looking over some seesaw experiments I was working on a while back. One thing that I did for the "standardDelay" was to use a functional interface ... I think I did it this way because I wanted flexibility to switch between busy-wait vs sleep (I've noticed sometimes when dealing with microseconds that is time.Sleep is not always super precise on every target, so that flexibility is useful to me).
Not saying that decision was correct/best (the code might not even be working, I see I still have some debugging print statements in there), but here is an example of what I mean in case you are curious about it:
https://github.com/bgould/go-seesaw/blob/main/seesaw.go#L13-L21
https://github.com/bgould/go-seesaw/blob/main/mini-tft-wing.go#L115-L119
There was a problem hiding this comment.
Good though. I've actually not considered that.
I've now simplified it further to leave more room open to also add a ReadWithWaiter(...) or similar in the future while simplifying the interface for the most common use cases.
The 'delay' argument felt a bit awkward anyways.
I think as a first start the current implementation will fit most of the applications.
If one must, one can change the dev.ReadDelay = ... before a read call.
|
This looks generally good to me, I will look over some more tomorrow. Was trying to find if I have my TFT Featherwing with me to test, but I think I might not have it (also will check more tomorrow). I can picture that maybe in addition to this low-level package, eventually there could be sub-packages for devices that use it, like /drivers/seesaw/mini-tft-featherwing. A "device" implemented on top of this package could look something like this: https://github.com/bgould/go-seesaw/blob/main/mini-tft-wing.go associated example: https://github.com/bgould/go-seesaw/blob/main/examples/mini-tft-wing/main.go (like I said I will try to see if I still have that board so that I can give your package a test drive) |
Indeed! And actually that's what I've done here: #564 I added ontop Keypad & NeoPixel and then wrapped it all to make the NeoTrellis work nicely :) Though in hindsight I think those are not yet polished enough for the tinygo-org/drivers repo ;) |
|
Bump? Any chance for a review? |
|
What's missing? Any interest? |
|
Sorry I was trying to find a board with seesaw on it to test, but I could not find. All in all, LGTM, now merging. Thank you for your contribution @trichner |
Ahh, that would have been great indeed. Thanks for merging. |
Here we go again. Simple driver for the "i2c to anything" seesaw chip.
https://learn.adafruit.com/adafruit-seesaw-atsamd09-breakout/overview
This also addresses half of #564.