This mixes some manual work (IWYU pragmas, a better solution to the Arduino
preprocessor macros problem) with automated running of the tools. At this
point, it would be too much work to separate these into distinct commits, and
there isn't that much value to doing so.
There are still some things we could do to make things more robust, as some of
the headers need to be in a certain order, which happens to be in the same sort
order used by IWYU (`testing/*` files need to come after certain headers than
include `Arduino.h`), but it's probably not worth the clutter of adding an `#if
1` just to stop IWYU from re-ordering them.
I tried to get `#pragma push_macro("max")/pop_macro("max")` to work, but ended
up getting completely nonsensical compilation errors, so I gave up on it.
Signed-off-by: Michael Richters <gedankenexperimenter@gmail.com>
Instead of using the phony `DEFAULT_GOAL` target, use make's special variable
`.DEFAULT_GOAL` to work around the problem of including arduino-cli.mk at the
top of the sketch makefile.
Signed-off-by: Michael Richters <gedankenexperimenter@gmail.com>
There were sometimes two blank lines instead of one between makefile targets,
without any clear pattern. Now each one is separated from the next by one blank
line.
Signed-off-by: Michael Richters <gedankenexperimenter@gmail.com>
Instead of one long line declaring a lot of makefile targets as phony, include a
line like `.PHONY: <target>` immediately above each one. This makes the
makefile longer, but it's now obvious if the target you're looking at is phony.
Signed-off-by: Michael Richters <gedankenexperimenter@gmail.com>
Now that it's not a `static` class function, we need to use a different
invocation, and we can't declare `isStickableDefault()` with the `always_inline`
attribute, or the user's override won't be able to call it because there will be
nothing to link to.
Signed-off-by: Michael Richters <gedankenexperimenter@gmail.com>
This testcase is for a version of TapDance that distinguishes between a "tap"
timeout and a "hold" timeout, allowing for two different `Key` values for the
same tap count.
Signed-off-by: Michael Richters <gedankenexperimenter@gmail.com>
This is closer to a real-world scenario than the QueueLeaker plugin testcase.
It doesn't test the bug as deliberately, but it shows the failure of Qukeys
prior to the fix.
Signed-off-by: Michael Richters <gedankenexperimenter@gmail.com>
`KeyAddrEventQueue::remove()` fails to confirm that the current queue is empty
before decrementing the length and shifting entries in the queue arrays. If
`remove()` or `shift()` is called when the queue is empty, `length_` gets
decremented from 0 to 255 (because it's unsigned), and then a large section of
memory gets shifted, mostly out of bounds of the event queue arrays, and
probably wreaking havoc with any number of things.
The plugin in this testcase should trigger this bug, and is detectable because
it affects the value for the current time. It's not guaranteed to detect this
bug, but it seems to be fairly consistent.
Signed-off-by: Michael Richters <gedankenexperimenter@gmail.com>
This is a demonstration that another plugin can use the new public OneShot
methods to turn a non-OneShot key into a OneShot.
Signed-off-by: Michael Richters <gedankenexperimenter@gmail.com>
OneShot keys should apply to all the key events generated by a Macros key, not
just the first one, even if the Macros key is injected by TapDance.
Signed-off-by: Michael Richters <gedankenexperimenter@gmail.com>
Michael Richters
Jesse Vincent
Florian Bruhin
Gergely Nagy