Struct ignore::WalkParallel
source · pub struct WalkParallel { /* private fields */ }
Expand description
WalkParallel is a parallel recursive directory iterator over files paths in one or more directories.
Only file and directory paths matching the rules are returned. By default,
ignore files like .gitignore
are respected. The precise matching rules
and precedence is explained in the documentation for WalkBuilder
.
Unlike Walk
, this uses multiple threads for traversing a directory.
Implementations§
source§impl WalkParallel
impl WalkParallel
sourcepub fn run<'s, F>(self, mkf: F)where
F: FnMut() -> Box<dyn FnMut(Result<DirEntry, Error>) -> WalkState + Send + 's>,
pub fn run<'s, F>(self, mkf: F)where F: FnMut() -> Box<dyn FnMut(Result<DirEntry, Error>) -> WalkState + Send + 's>,
Execute the parallel recursive directory iterator. mkf
is called
for each thread used for iteration. The function produced by mkf
is then in turn called for each visited file path.
sourcepub fn visit(self, builder: &mut dyn ParallelVisitorBuilder<'_>)
pub fn visit(self, builder: &mut dyn ParallelVisitorBuilder<'_>)
Execute the parallel recursive directory iterator using a custom visitor.
The builder given is used to construct a visitor for every thread used by this traversal. The visitor returned from each builder is then called for every directory entry seen by that thread.
Typically, creating a custom visitor is useful if you need to perform
some kind of cleanup once traversal is finished. This can be achieved
by implementing Drop
for your builder (or for your visitor, if you
want to execute cleanup for every thread that is launched).
For example, each visitor might build up a data structure of results corresponding to the directory entries seen for each thread. Since each visitor runs on only one thread, this build-up can be done without synchronization. Then, once traversal is complete, all of the results can be merged together into a single data structure.