Why does pg.Cursor still use callbacks?

[Bluenix2]

Looking at the the pg.Cursor API documentation it appears to be using callbacks, is there a reason that there hasn't been a switch to the Promises API introduced in ES6?

Other parts of the package shows the usage of promises. As can be seen here.

[sehrope]

All of that code predates Promises by many years and nobody has gotten around to adding it. It's common for anybody using the Cursor API in modern applications to add their own Promise wrapper. Ex: https://github.com/sehrope/postgres-with-nodejs-code/blob/85fd7acf5b4e1be5a124bbf39c68bbc2e5591489/src/00026-memory-usage-with-cursor.js#L5-L15

I think it'd be fine to add it directly to the Cursor.read(...) function. It could mirror the query API and check if the second parameter is not defined, and if so, return the result as a Promise. The close function is a bit trickier as right now the callback is not required so changing it to return a Promise (when there's no callback) could lead to a dangling Promise in existing code.

[Bluenix2]

Is there a reason that Cursor doesn't use ES6 classes or can I make that switch as well?