In all honesty, I don't have a tonne of experience with setting up code coverage in non-Golang environments.

I think that https://coveralls.io looks very nice and has a good interface. That said, I'm always a little bit wary to adopt SaaS when there could be a more "long lived" solution. (for example, choosing astyle over clang-format despite astyle existing on Sourceforge)

I was recommended gcov as well. I'm probably going to read the manual for gcov and then give it a go. If anyone has firm opinions and/or recommendations, I would be very interested to hear them.

Seems like a lot of C test coverage tools are proprietary.

Well, this is a bit later...

(included image is of lcov being run to generate code coverage)

Turns out it wasn't actually that hard. I was just being kind of daft.

I'm not 100% sure this actually makes 100% sense, so please do check me if something this diagnostic is showing seems inaccurate.

It seems like coveralls has a nice interface, but it's also just using gcov underneath.

So much for a diverse ecosystem of C code coverage tools! :)

Thanks for revisiting this. Is this output for a single execution of test_kem for a BIKE algorithm? Is it able to aggregate coverage across multiple runs?

Yeah it appears to be able to do a decent job of aggregating across multiple tests being run. I just need to figure out how to run all the tests haha.

also works for AES.

Seems to use pytest if I'm not wrong.

Alright, looks like this is the approximate flow:

# build with --coverage flag
cmake -DCMAKE_BUILD_TYPE=Debug -DCMAKE_C_FLAGS="--coverage" -DCMAKE_CXX_FLAGS="--coverage" ..

# run files
<trying to sort this bit out>

# use lcov to automate running gcov for us
lcov --capture --directory . --output-file coverage.info

# generate an HTML report
genhtml coverage.info --output-directory coverage_html

Hi @aidenfoxivey! I took a stab at this a couple of months ago but haven't had time to properly see it through. Maybe my branch would be helpful to you?

That's awesome. Thanks @SWilson4. I'll give that a look.