Trait idmap::table::IdTable[−][src]

pub trait IdTable: Debug + Clone {
    fn new() -> Self;
    fn with_capacity(capacity: usize) -> Self;
    fn get(&self, key_index: TableIndex) -> Option<TableIndex>;
    fn create_mut(&mut self, key_index: TableIndex) -> &mut TableIndex;
    fn set_raw(&mut self, key_index: TableIndex, value: TableIndex);
    fn clear(&mut self);
    fn reserve(&mut self, amount: usize);
}

Expand description

Assigns unique u32 indexes to each IntegerId key, which can be used to modify the behavior and performance of a DenseEntryTable.

Currently the only implementation is OrderedIdTable, which preserves ordering by using a Vec

DenseEntryTables need a separate IdTable since entries will be stored densely, though a DirectEntryTable doesn’t need one at all.

Required methods

[src]

fn new() -> Self

Create a new table

[src]

fn with_capacity(capacity: usize) -> Self

Create a table, initialized with the specified capacity

[src]

fn get(&self, key_index: TableIndex) -> Option<TableIndex>

Determine the index of the specified key, which may or may not be valid.

In other words this may have ‘false positives’, so that Some(TableIndex::from(key)) is a correct implementation even though the index isn’t necessarily in-bounds for the entry table. However, it can never have ‘false negatives’ so None and Some(TableIndex::INVALID) are both wrong.

The TableIndex::INVALID index is special case and should never be considered valid under any circumstances. It is logically equivalent to a None result, but kept for performance reasons.

[src]