The center for all discussion and news regarding C++.
tokei is the tool that came to mind reading your title.
Getting fancier info depends on the fanciness of the tokenizer used. Maximum fanciness probably requires a full language parser. Info beyond code/tokens requires more than just a parser (language server territory).
It might be possible to get this using treesitter, though not sure
I'm unsure if that's even possible in C++ for a "normal program" instead of requiring a full-fledged compiler and even maybe a linker, as C++ syntax is completely context-dependent. There's no way to tell if i+=b; increments a value or is actually a dynamic hierarchy cast that prints a message to screen and invokes a vector copy with data downloaded from the internet.
I'd venture there's something in Clang that does this? (clang-tidy, clang-format, who knows), since they implement the syntax tree parser thing.
I didn't even think about how context dependent C++ is until you mentioned that. That's a really good point
Yeah basically C++ is fun (or "fun") in that in order to try and get what code does or what it means you have pretty much no other alternative than trying to compile it.
There are a number of terms that have only one meaning and can appear in only one "semantic place", such as namespace, return and... I think that's it? Maybe (maaaaybe) .*. So it's little enough that doesn't give you that much good information if you try and hand-parse them out from the code you have.
This is a lot more complicated than I first anticipated. I guess its probably not a common tool then 😅
Whch is why the cool thing(TM) nowadays is to use an IDE that has Clang integration, and let the compiler that already compiles your code anyway, do all that sweet statistics job for you.
Maybe count the size of the output of the compiler in bytes? It has done all the work already…
That would only tell me how large each object/asm file is, its not really indicative of the lines of code as it could expand to quite a large size once compiled.
I'm looking for more of an analysis of what has been written and its ratio to the line count.
Thanks for suggesting this, but its not quite what I'm looking for.
I wonder if it would be possible to build such a tool on top of tree-sitter (although not sure tree-sitter's C++ grammar can handle modules yet)
Probably something written in perl.