getting type name at compiletime

[cobalt herald]

pls review
https://godbolt.org/z/P5TzWoabY

[hollow python]

Instead of return ""; you might want to #error for an unsupported compiler

Depending on your design goals / what guarantees your library wants to provide

[cobalt herald]

alright

[hollow python]

My implementation of a type_name utility is here if it's of any help https://github.com/jeremy-rifkin/libassert/blob/8941db76468e0c57600494cef67b13c2d4672581/include/assert.hpp#L557

[cobalt herald]

oo

[cobalt herald]

i don't have any substring functions like that yet, do you think i should use signature.rfind('>') - start or signature.size() - start - 7 for the msvc thing?

oh the function is literally defined above

[hollow python]

I found the substring helper a nice way to be expressive but you can definitely do it with signature.rfind('>') - start

[cobalt herald]

commencing thievery

[cobalt herald]

so i should prefer that over hardcoding 7 i see

your helper doesnt handle i being npos

does it happen anyway? HM

no

i dont think

[cobalt herald]

i made before, after, and between string functions

[cobalt herald]

@hollow python https://github.com/Eczbek/xieite/blob/main/include/xieite/types/name.hpp

is this the only way to achieve this?

[hollow python]

There's never just one way to do a problem but that should work

[cobalt herald]

ok is there a better way?

[hollow python]

What are your goals?

Rediculous efficiency? Clean and expressive code?

[cobalt herald]

best practices

[hollow python]

One thing that would help are comments explaining what __PRETTY_FUNCTION__/__FUNCSIG__ look like on the various compilers

And in turn explaining what on earth is going on with between("= ", "]")

[cobalt herald]

ah

ok