Storage slot hash calculation

Sway
1

The storage slots for each storage variable that is defined in storage block, was created by hashing the storage_X string where X was the integer value starting from 0 based on the order of the variable in the storage block.

I noticed that after upgrading from 0.60.0 to 0.63.1 this is not the case anymore.

I have this simple contract:

contract;

storage {
    variable: u64 = 0,
}

abi MyContract {
    fn test_function() -> bool;
}

impl MyContract for Contract {
    fn test_function() -> bool {
        true
    }
}

and when I compile it the following slots.json gets generated:

[
  {
    "key": "7cee3037d9b2fdd229ca5800cdd80a341eaac931ca6d9619736c34428a54bf43",
    "value": "0000000000000000000000000000000000000000000000000000000000000000"
  }
]

which is not matching the sha256 value I would expect:

$ echo -n "storage_0" | shasum -a 256 -
f383b0ce51358be57daa3b725fe44acdb2d880604e367199080b4379c41bb6ed  -

Could someone please provide an explanation how is the slot value calculated now, or provide a link to the relevant source code?

2

Ok, I found the relevant PR here: Storage name based slots and namespaces. by esdrubal · Pull Request #6064 · FuelLabs/sway · GitHub

It is clear how the slot value is computed, but not clear how the field_id is. The slot is based on the following string:
storage::<storage_namespace_name1>::<storage_namespace_name2>.<storage_field_name>

But the field_id (as documented in the PR) based on this:
storage::<storage_namespace_name1>::<storage_namespace_name2>.<storage_field_name>.<struct_field_name1>.<struct_field_name2>

It is not clear what the struct_field_name is representing here.

When I tried to list both the slot and field_id of my storage test contract above, they have the same value, which looks like the struct_field_name (whatever that is) is an empty string.

I also tried to create a struct variable in storage with some fields, but got the same result: slot == field_id.