i would like to add tests for this, but the test suite fails for me locally even without any changes :/ so it would be kind of difficult for me to make sure my new test passes.

chord-double.t: PASSED
WARNING: Insufficient output, timing out one more time...
chord-hold.t: ERROR: Extraneous keys.
Input:
a down
b down
199ms
x down
x up
a up
b up
1ms
a down
b down
200ms
x down
x up
a up
b up

Expected Output:     Output:
a down               a down
b down               b down
x down               x down
x up                 x up
a up                 a up
b up                 b up
control down         a down
x down               b down
x up                 x down
control up           x up
                     a up
                     b up

Should a mention be made in the manpage?

good call! done :)

Apologies for the delay. I had earmarked this for reply and have not had a chance to address it.

Can you elaborate on the use case? Mouse support is currently highly experimental (many things don't work properly), which is why * currently only captures keyboards. Devices like trackpads are largely unsupported (and will misbehave if captured by keyd), which is why the feature currently requires the user to explicitly add their pointing device to the config file. Most people only have one or two pointing devices, and a wildcard which matches all mice might cause some nasty surprises, so I am generally opposed to it.

i like my configuration files to be portable across machines. not having a wildcard means that either my config is specific to the exact hardware plugged in, or i have to do runtime detection in my install script.

i understand that mouse support is experimental, but, well, no one has to use the new wildcard. i think keeping the existing notice that mouses might break is enough, i don’t think banning mouse wildcards would discourage someone who’s already decided they want this.

i like my configuration files to be portable across machines.
[...]
i understand that mouse support is experimental, but, well, no one has to use the new wildcard. i think keeping the existing notice that mouses might break is enough, i don’t think banning mouse wildcards would discourage someone who’s already decided they want this.

Real portability is not achievable by dint of the fact that mouse functionality will break at least 50% of the time. If you just copy a config containing this wildcard, it has the potential to render the machine unusable. The user should be explicitly whitelisting mice which are known to work.