Sigh. Wait until she learns about the static initialization order fiasco...

She knows. Code where it matters is wrong.

Sounds like a have to C it to believe it.

I've written code like this in C to create library functions that will be linked in to other projects, but where there is no 'main' function in the C file. Do people use C++ the same way, or is it best practice to include 'main?'

It's the same as in C. The observation here is that variables with static storage duration can be initialized by running user code before main() begins.

Oh! Thank you. That makes much more sense now.

It’s also the same in .NET. If your Main() function references types that have a static constructor, the constructors could run before Main().

It’s a surprising thing when one first discovers it, but it makes a lot of sense.

It didn't crash until after its work was done. It could even have trapped the segfault and exited cleanly.

Couldn't just call exit(0)?

Maybe, or _exit().

A more useful application of this same idea ("you can do stuff during static initialization") is for building a static registry.

https://artificial-mind.net/blog/2020/10/17/static-registrat... gives a good overview.

[1] p.99 information might be of interest

   cc -Wl,-m main.c > main.linker.map

related search terms: _crtStartUp crt0 crt1

[1] : https://progforperf.github.io/Expert_C_Programming.pdf

Is this what passes for "technical know-how" nowadays?

Just read Hongjiu Lu's classic paper ELF: From The Programmer's Perspective to get an idea of how everything comes together in C/C++ land.

But in today's non-programmer post-covid world, ELF is seasonal, COFF is year round.

edit: COFF phoneticly eqivalent to cough.