< elichai2>
Another reason why we want enum classes is that you can't do forward declaration with a "regular" enum :/
< hebasto>
elichai2: from EMC++ item 10 - "In C++11, unscoped enums may also be forward-declared, but only after a bit of additional work."
< elichai2>
lol. what is "but only after a bit of additional work"? I couldn't make it work(and neither people in #c++) even though some stackoverflow people said it can be done
< ryanofsky>
you can forward declare normal enums with "enum MyEnum : int;" You need to explicitly specify the integer type so the compiler knows what type will fit all the enum values
< elichai2>
ryanofsky: don't you also need to explictly specify the integer type on the enum definition?
< elichai2>
well IWYU doesn't agree with me so I guess i'm wrong. ha. (anyone understands this?)
< elichai2>
god we have a lot of files lol
< ryanofsky>
elichai2, yes declaration and definition should be consistent and both use the same types or no types. but the compiler can't detect inconsistencies in all cases (different .cpp files)
< bitcoin-git>
[bitcoin] IPGlider opened pull request #17884: utils: Use fallback for macOS in AllocateFileRange (master...workaround_apfs_allocation_problem) https://github.com/bitcoin/bitcoin/pull/17884
< ariard>
what the rational to update protocol version number ? #16442 adds new p2p messages, but don't bump it, at the contrary protocol version has been bumped for bip37/bip152