Struct EvolutionAnswers 

#[non_exhaustive]
pub struct EvolutionAnswers {
    pub parameterized_string_widening_is_lossless: bool,
    pub nested_field_addition_is_lossless: bool,
}

What this engine keeps through a change to a table that already holds rows.

Every answer defaults to false, which means nothing has been measured rather than this engine loses data. Razeri’s classifier treats a change it cannot prove keeps every value as one that does not, so a silence here costs an operator a warning they can read and never costs them a row.

An answer only ever makes a verdict gentler. Razeri proves what it can from the types themselves first; these answers are asked afterwards, about the changes the engine’s own spelling left unprovable.

Fields (Non-exhaustive)

Non-exhaustive structs could have additional fields added in future. Therefore, non-exhaustive structs cannot be constructed in external crates using the traditional Struct { .. } syntax; cannot be matched against without a wildcard ..; and struct update syntax will not work.

parameterized_string_widening_is_lossless: bool

This engine widens a string’s declared length in place, keeping every value.

A column declared varchar(10) and now declared varchar(40) holds a strict superset of what it held, as a type. Whether the engine gets it there without rewriting the column is the engine’s own business, and this is where it says so.

The question is asked at all because a length is spelled differently on every engine — VARCHAR(40) on one, STRING(40) on another, NVARCHAR(40) on a third — and the length in a spelling Razeri does not write is a length Razeri cannot read back. A true here is the dialect saying “that pair is a widening on my engine, and nothing is lost”.

nested_field_addition_is_lossless: bool

This engine adds a field to a record a table already holds, keeping every value.

The change is a record column that gained a field and changed in no other way. Some engines take it as a change to the column’s declared shape and keep every row; others have no way to make it at all, and the column has to be rebuilt.

false is the careful answer, and it is the right one for every engine that has no records to begin with.

Implementations

impl EvolutionAnswers

pub const fn none() -> Self

Nothing measured — the answer for an engine nobody has asked.

pub const fn widening_a_string_in_place() -> Self

This engine widens a string’s declared length in place. See EvolutionAnswers::parameterized_string_widening_is_lossless.

pub const fn and_a_nested_field(self) -> Self

The same answers with the nested-field one added, so an engine that keeps both states both without a constructor per pair.

Trait Implementations

impl Clone for EvolutionAnswers

fn clone(&self) -> EvolutionAnswers

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Copy for EvolutionAnswers

impl Debug for EvolutionAnswers

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for EvolutionAnswers

fn default() -> EvolutionAnswers

Returns the “default value” for a type.

impl Eq for EvolutionAnswers

impl PartialEq for EvolutionAnswers

fn eq(&self, other: &EvolutionAnswers) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for EvolutionAnswers