How does "tablename"_n differ to N(tablename) when defining a multi_index?

Question

I've seen two examples of how to define the table name of multi_index:

• N(tablename)
• "tablename"_n (source)

Initially I tried using the former, but resulted in the below compilation error. I have since been using the latter without any issue. Are both supported or was this changed in a subsequent release?

error: use of undeclared identifier 'tablename'
        typedef eosio::multi_index<N(tablename), tablerow> tablerow_index;

For reference, I've been using EOSEasyContract for compilation, I'm not sure if this makes any difference. My understanding is it should as it's using a Docker container to run the compilation.

Answer 1

The N macro was replaced with the _n operator. It is one of the breaking changes introduced recently. Here is the complete list, https://github.com/EOSIO/eosio.cdt/blob/master/README.md

Removal of the N macro. The ""_n operator or the name constructor should be used as a type safe replacement. Example: N(foo) -> "foo"_n, or N(foo) -> name("foo").