Struct FsyncComponents 

pub struct FsyncComponents { /* private fields */ }

What: The components one core.fsync value makes Git flush. Why: One value type for both questions, so the index writer and the import step cannot disagree about how Git reads the setting.

In TS you’d write (pseudocode):

type FsyncComponents = { bits: number };

Implementations

impl FsyncComponents

Reading and asking.

pub fn default_set() -> Self

What: The set Git flushes when core.fsync is not set. Why: An unset key is the default, which flushes packs and their metadata but not loose objects, references or the index.

In TS you’d write (pseudocode):

FsyncComponents.default()

pub fn parse(value: &str) -> Self

What: The set one core.fsync value makes Git flush. Why: The value starts again from the default at each entry. In one value none (only as the whole rest of the value) clears the set, a word with a - is removed and any other is added, a word names every component whose name it is a prefix of (ind is index, pack also names pack-metadata), a word that names none is ignored, and the result is the set without the removed components plus the added ones, so an added component stays even when it is also removed.

In TS you’d write (pseudocode):

FsyncComponents.parse('added,-pack')

pub fn names_loose_object(self) -> bool

Whether loose objects are flushed.

pub fn names_index(self) -> bool

Whether the index file is flushed.

pub fn names_pack(self) -> bool

Whether a pack file or a pack’s metadata is flushed.

pub fn names_pack_and_metadata(self) -> bool

Whether both a pack file and a pack’s metadata are flushed.

pub fn names_reference(self) -> bool

Whether the files of the reference store are flushed.

pub fn without_packs(self) -> Self

What: This set without the pack file and the pack’s metadata. Why: The import step of a landing writes a temporary pack that it deletes when the batch is small, and native Git writes loose objects, so a repository that does not flush loose objects gets no pack flush for it.

In TS you’d write (pseudocode):

set.without('pack', 'pack-metadata')

pub fn with_packs(self) -> Self

What: This set with the pack file and the pack’s metadata added. Why: Where loose objects are flushed, a pack that the import keeps instead of unpacking holds the same objects and is flushed with its index.

In TS you’d write (pseudocode):

set.with('pack', 'pack-metadata')

pub fn value(self) -> String

What: A core.fsync value that makes Git flush exactly this set. Why: Each word is spelled in full, because a word is a prefix match (pack alone would also name pack-metadata), and none ends the value, which starts the set empty. A component that no single word names (the mapping between hash algorithms) is left out.

In TS you’d write (pseudocode):

set.toValue() // 'commit-graph,reference,none'

Trait Implementations

impl Clone for FsyncComponents

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for FsyncComponents

impl Debug for FsyncComponents

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

Formats the value using the given formatter.

impl Eq for FsyncComponents

impl PartialEq for FsyncComponents

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

Equality operator ==.

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

Inequality operator !=.

impl StructuralPartialEq for FsyncComponents