sum types should have a default 0 value for the first type mentioned in the list of types. about v HOT 4 CLOSED

One problem that will have to be solved too, when the first variant type of a sumtype declaration, gets used to define the default value, is that currently v fmt -w . will reorder the sumtype variants in alphabetical order, and it should not, since the first type will have a semantic meaning in the future.

from v.

The same is true for interface values too - they also do not have default values.

For them, the problem is mitigated a bit because we do have an error for some situations involving them:

a.v:2:6: error: cannot instantiate interface `Abc`
    1 | interface Abc{}
    2 | a := Abc{}
      |      ~~~~~
    3 | dump(a)
#1 20:03:14 ^ master /v/oo>

but others are still allowed:

a.v:2:6: warning: arrays of interfaces need to be initialized right away, therefore `len:` cannot be used (unless inside `unsafe`, or if you also use `init:`)
    1 | interface Abc{}
    2 | a := []Abc{len:4}
      |      ~~~~~~
    3 | dump(a)
[a.v:3] a: [unknown interface value, unknown interface value, unknown interface value, unknown interface value]

and even runtime errors, with no complains from the compiler:

interface Abc{}
struct MyStruct { i Abc }
a := []MyStruct{len:4}
dump(a)

producing:

#148 20:05:19 ^ master /v/oo>v run a.v
/tmp/v_1000/a.01J516DMX1F1X3GBBFX1R951GE.tmp.c:2277: at indent_main__Abc_str: RUNTIME ERROR: invalid memory access
/tmp/v_1000/a.01J516DMX1F1X3GBBFX1R951GE.tmp.c:2264: by indent_main__MyStruct_str
/tmp/v_1000/a.01J516DMX1F1X3GBBFX1R951GE.tmp.c:2216: by indent_Array_main__MyStruct_str
/tmp/v_1000/a.01J516DMX1F1X3GBBFX1R951GE.tmp.c:2210: by Array_main__MyStruct_str
/tmp/v_1000/a.01J516DMX1F1X3GBBFX1R951GE.tmp.c:2233: by _v_dump_expr_Array_main__MyStruct
/tmp/v_1000/a.01J516DMX1F1X3GBBFX1R951GE.tmp.c:13400: by main__main

from v.

It could always alphabetize all the types after the first, but that seems like more trouble than it's worth.

Changing the defaults for interfaces may also need it's own separate issue... seems like a much bigger problem than sum types.

from v.

Just wanted to confirm that agree with this notion.

from v.