C++ Modules doesn't count as a part of C++ source code

[SpidX9226]

Duplicates

• I have searched the existing issues

Current behavior 😯

Tokei already supports C++ modules, but onefetch does not display them as part of the C++ source code.

Expected behavior 🤔

Onefetch should include C++ Modules in C++ statistic.

Steps to reproduce 🕹

1. Create any .ixx/.cppm/.cxxm/.c++m/.ccm file
2. Done

Additional context/Screenshots 🔦

As we can see here:

Possible Solution 💡

No response

[spenserblack]

Since Tokei counts C++ modules as its own language, I think we can consider this a new language request.

IMO the C++ logo can just be copied and pasted for this one.

[SpidX9226]

Since Tokei counts C++ modules as its own language, I think we can consider this a new language request.

IMO the C++ logo can just be copied and pasted for this one.

Tokei also counts header files as a separate language, whereas onefetch doesn't; therefore, in my opinion, it would be ambiguous to consider modules a language but not header files.

[ryux1]

I can implement the collaborator-suggested representation as a separate CppModule language entry, reusing the existing C++ logo, colors, chip, and icon. That follows Tokei’s current CppModule classification and makes .cppm, .ixx, .ccm, .mpp, .mxx, .cxxm, .hppm, and .hxxm visible instead of dropping them. I’ll keep C++ headers out of scope for this issue.

[spenserblack]

Tokei also counts header files as a separate language, whereas onefetch doesn't

I don't think we ever implemented any logic to combine the Cpp and CppHeader variants into one. C++ headers don't show up as a separate language because they don't show up at all; we never registered the CppHeader variant. Here's a quick and lazy check:

# Language name to lines-of-code map
languages = {
    "CMake": 26,
    "C++": 34,
    "C++ Header": 17,
    "C++ Module": 35,
    "JSON": 24,
}

def ratio_of(combined_language_counts: list[str], all_languages: list[str]):
    return sum(languages[key] for key in combined_language_counts) / sum(languages[key] for key in all_languages)

print("Assuming C++ Header is registered and combined with C++:", ratio_of(["C++", "C++ Header"], ["CMake", "C++", "C++ Header", "JSON"]))
print("C++ Header is not registered and isn't factored into the C++ count nor the overall total:", ratio_of(["C++"], ["CMake", "C++", "JSON"]))

Assuming C++ Header is registered and combined with C++: 0.504950495049505
C++ Header is not registered and isn't factored into the C++ count nor the overall total: 0.40476190476190477

The second output, where C++ Headers aren't counted at all, matches the stats that are currently reported.

However, this issue gave me an idea: perhaps languages in onefetch could have an optional parent language entry. If a language has a parent, we map all of its stats to that variant. So C++ Headers and C++ Modules can have C++ as its parent.

[be-student]

I’ll take this. I’ll map Tokei’s C++ module variant into Onefetch’s existing C++ language, aggregate both LOC buckets, and add coverage for detection and summation.

[ryux1]

I had prepared the separate-language version locally, but the parent-language aggregation direction changes the shape of the fix. Since you’re taking that approach, go ahead—I’ll step back so we don’t duplicate the work.

[v01dst]

I'd like to pick this up — could you assign me? I'll follow the contribution guidelines and open a PR referencing this issue shortly.

[spenserblack]

I'd like to pick this up — could you assign me?

#1856 already exists.