Keeley Hoek comments

Results 20 comments of


                                            Keeley Hoek

The `pki.decryptRsaPrivateKey` function can return null if the password is wrong

This tripped me up, too

Misses in coverage

I have a whole bunch of `match self.blah { }`s where the line containing the `match` itself is marked not-covered.

`conv {to_rhs, ...}` behaves poorly with `nat.modeq`.

Presumably the idea in the implementation is that `is_relation` should fail when you pass it a ternary (n>2-ary) function?

Range types for integers (or refinement types?)

Is anything like this in the language yet?

NKRO breaks layer-tap with "extended" keycodes

Thanks for your help! The thing is, from what I can see this behavior is happening on both Windows and Ubuntu, and doesn't happen for the < 12 function keys...

NKRO breaks layer-tap with "extended" keycodes

Yes, after setting up wireshark for USB capture I can confirm that a single tap of my `&lt my_layer F24` key does cause 2 USB packets to be sent (press...

NKRO breaks layer-tap with "extended" keycodes

Another thing is that, even more rarely, once I get the "right combination" of spammy presses to get the `F24` to appear tapping the key will successfully emit more `F24`s...

NKRO breaks layer-tap with "extended" keycodes

OK, I believe this is a bug in upstream Zephyr. By inserting print statements I see ZMK making a call to `hid_int_ep_write()` passing the 22-byte string `01 00 00 00...

NKRO breaks layer-tap with "extended" keycodes

Okay, found the problem. There is a data race in `usb_hid.c` on many platforms; the call to `hid_int_ep_write()`, a Zephyr function, is not in general synchronous and doesn't always copy...

NKRO breaks layer-tap with "extended" keycodes

The fix is in #2257. Tested working on a real Adv360.