Rework kernel error categories

[PhilippGackstatter]

What should be done?

We should consider reworking how transaction kernel errors are structured on a category level. Category here means account, prologue, epilogue, faucet, ... These errors are already grouped in tx_kernel_errors.rs:

pub const ERR_ACCOUNT_CODE_COMMITMENT_MISMATCH: u32 = 0x0002004c;
pub const ERR_ACCOUNT_CODE_IS_NOT_UPDATABLE: u32 = 0x0002003d;
pub const ERR_ACCOUNT_...

pub const ERR_FUNGIBLE_ASSET_AMOUNT_EXCEEDS_MAX_ALLOWED_AMOUNT: u32 = 0x00020042;
pub const ERR_FUNGIBLE_ASSET_DISTRIBUTE_WOULD_CAUSE_MAX_SUPPLY_TO_BE_EXCEEDED: u32 = 0x00020021;
pub const ERR_FUNGIBLE_ASSET_...

Here they are grouped by the category in the name but the error code does not reflect the categories.

How should it be done?

It would be nice to be able to programmatically determine the category of an error from the error code, similar to how one can determine from an HTTP status code (1xx, 2xx, 3xx, ...) the category (Informational, Success, Redirect, ...).

For that, we might want to come up with ranges for categories, for example:

• ERR_ACCOUNT_* might be within 0x00020040 - 0x00020050
• ERR_FUNGIBLE_ASSET_* within 0x00020050 - 0x00020060
• and so on...

The task is to figure out what range size makes sense and then rewrite the error codes to be within those categories.

When is this task done?

When the errors can be categorized from their code.

Additional context

For reference, the original context where this came up.