could use some help with an elementary example

[VictorEijkhout]

It looks like all your examples are for SI. Can you help me get this going:

#include <mp-units/systems/si.h>
using namespace mp_units;
inline constexpr struct person final : named_unit<"p"> {} person;

int main( int argc,char const * const * argv ) {

  auto classroom  = 30 * person;

[mpusz]

Sure! 😃

However, we need more info first. How do you want to model persons? Is it a dimensionless quantity? If so, then it means that a unit one will also work fine for it, and the code below would compile:

auto classroom = 30 * person;
classroom = 29 * one;

If the above is incorrect, you should probably make the count of persons a distinct quantity with its own unit. If that is the case, you may model it with:

#include <mp-units/core.h>
#include <iostream>

using namespace mp_units;

inline constexpr struct persons_count_dim final : base_dimension<"P"> {} persons_count_dim;
inline constexpr struct persons_count final : quantity_spec<persons_count_dim> {} persons_count;

inline constexpr struct person final : named_unit<"p", kind_of<persons_count>> {} person;

int main()
{
  quantity classroom = 30 * person;
  std::cout << classroom << "\n";
}

https://godbolt.org/z/96jnaxcWP

The above can be extended to define quantity types for counts of students and teachers to make them strong, not interconvertible types.

[VictorEijkhout]

Interesting. The core.h file is not installed by the install rule.

[staff src:26] find .. -name CMakeLists.txt -exec grep core.h {} \; -print
[staff src:27]

[staff src:30] find .. -name CMakeLists.txt -exec grep core.*h {} \; -print
            include/mp-units/bits/core_gmf.h
../src/core/CMakeLists.txt

Btw, since I don't use conan, I do make install in the build directory. That ought to work since cmake makes perfect makefiles, not?

[mpusz]

Interesting. The core.h file is not installed by the install rule.

That is strange indeed, as it gets installed in our CI runs:

• https://github.com/mpusz/mp-units/actions/runs/14264342581/job/39982675401#step:16:411
• https://github.com/mpusz/mp-units/actions/runs/14264342562/job/39982677910#step:20:21

Could you please double-check that?

[mpusz]

What about the proposed solution? Is it what you expected?