Question about terminology: `make` vs `new` vs `create`

Explain
1

I remember I’ve asked about this about a year ago on discord and gingerBill had answered my question, but since then I’ve forgotten the exact answer.

In the Odin’s core library there are a lot of places where objects are being created or initialized. From what I’ve observed the naming conventions come in pairs:

  • new / free
  • make / delete
  • create / destroy
  • init / ???

Can someone clarify the exact differences between these pairs of naming conventions for procedures?

5 Likes
2

My conventions are not really about the pairs. Except for the attributes @(init) and @(fini), procedures do not necessarily have pairs.

  • new allocates a value of a certain type, not necessarily “constructed”
  • free deallocates/frees a rawptr that was previously allocated
  • make allocates the underlying data for a built-in data type and returns the result as an aggregate record (e.g. slice, dynamic array, map)
  • delete deallocates/frees the underlying data for a built-in data type (e.g. slice, dynamic array, map, string)
  • create will allocate and construct a certain type, and return a pointer
  • init will construct a certain value passed by pointer to it
  • destroy will deallocate and/or deconstruct a certain value passed to it (by value or by pointer)

So init is usually paired with destroy too.

13 Likes