NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
crates.io
crates.io, #3145 most downloaded on crates.io
Use `dua-core` instead
Latest stable
0.9.0
Last release 3 months ago
05 Aug 2026
Release timing varies
gaps range from 9 days to 3.6 years
Most releases are documented
notes for 12 of 14 stable releases
2 versions withdrawn
withdrawn after publishing
8 years old
16 releases · first in 2019
Drag across the ruler to compare, or tap a release to open it.
This crate is no longer maintained or supported. Use `dua-core` instead.
This crate is no longer maintained or supported. Use dua-core instead.
Clippy helped 1 time to make code idiomatic.
<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
gix-testtools to allow state reuse for directories and simpler testing (2ea7375)Your coding agent can read these notes before it upgrades. Set up the MCP server →
re-export rayon in the crate root. This makes creating a ThreadPool easier as it doesn't force us to maintain our own rayon dependency.
rayon in the crate root.
This makes creating a ThreadPool easier as it doesn't force us to
maintain our own rayon dependency.<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
rayon in the crate root. (b49e157)
</details>rayon in the crate root.
This makes creating a ThreadPool easier as it doesn't force us to
maintain our own rayon dependency.<csr-read-only-do-not-edit/>
<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
Parallelism::RayonExistingPool::busy_timeout is now optional. That way we can indicate that no waiting should be done as we know the given threadpool
Parallelism::RayonExistingPool::busy_timeout is now optional.
That way we can indicate that no waiting should be done as we know the
given threadpool has enough resources.<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
Parallelism::RayonExistingPool::busy_timeout is now optional. (3a71721)
</details>Parallelism::RayonExistingPool::busy_timeout is now optional.
That way we can indicate that no waiting should be done as we know the
given threadpool has enough resources.<csr-read-only-do-not-edit/>
<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
This release makes iterator creation fallible to avoid potential hangs when there is no available thread to process any of the iterator work.
This release makes iterator creation fallible to avoid potential hangs when there is no available thread to process any of the iterator work.
WalkDirGeneric::try_into_iter() for early error handling.
If we can't instantiate the iterator due to a busy thread-pool,
we can now abort early instead of yielding a fake-entry just to
show an error occurred. This is the preferred way to instantiate
a jwalk iterator.Detect possible deadlocks when instantiating a parallel iterator. Deadlocks can happen if the producer for results doesn't start as there is no free thread on the rayon pool, and the only way for it to become free is if the iterator produces results.
We now offer a busy_timeout in the relevant variants of the
Parallelism enumeration to allow controlling how long we will wait
until we abort with an error.
Clippy helped 1 time to make code idiomatic.
<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
WalkDirGeneric::try_into_iter() for early error handling. (7d5b8b8)stalling issue when threadpool is used is no more. The issue seems to have been that install blocks whereas spawn properly releases the main thread.
stalling issue when threadpool is used is no more.
The issue seems to have been that install blocks whereas spawn
properly releases the main thread.
This seems to have been changed subtly due to changes in rayon,
which breaks an assumption on how the code is executed.
Replacing install() with spawn calls resolved the issue.
<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
<csr-id-bd3e88017ea29c3b89b518f3a721ba35577b7666/> stalling issue when threadpool is used is no more.
The issue seems to have been that install blocks whereas spawn
properly releases the main thread.
This seems to have been changed subtly due to changes in rayon,
which breaks an assumption on how the code is executed.
Replacing install() with spawn calls resolved the issue.
<csr-read-only-do-not-edit/>
<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
The first release under new ownership with no user-facing changes.
The first release under new ownership with no user-facing changes.
cargo smart-release for releases.cargo clippy.Clippy helped 1 time to make code idiomatic.
<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
du example (0786beb)(cargo-release) jwalk- version 0.6.0
(cargo-release) jwalk- version 0.6.0
Added depth and path being read to params to ProcessReadDirFunction callback.
Allow setting initial root_read_dir_state (ReadDirState) instead of always getting ::default() value.
<csr-read-only-do-not-edit/>
<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
(cargo-release) jwalk- version 0.5.2
(cargo-release) jwalk- version 0.5.2
<csr-read-only-do-not-edit/>
<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
85ef009)ea66ef8)2459776)57860ff)a431561)fe12f26)64ddb05)32e46c6)3bd6618)b4776b9)d630f80)6d93359)
</details>(cargo-release) jwalk- version 0.5.1
(cargo-release) jwalk- version 0.5.1
<csr-read-only-do-not-edit/>
<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
(cargo-release) jwalk- version 0.5.0
(cargo-release) jwalk- version 0.5.0
<csr-id-11fa0bc9e8541af333aafb41cc89218435474df4/>
First major change is that API and behavior are now closer to [walkdir] and
jwalk now runs the majority of walkdirs tests.
Second major change is the walk can now be parameterized with a client state
type. This state can be manipulated from the process_read_dir callback and
then is passed down when reading descendens with the process_read_dir
callback.
Part of this second change is that preload_metadata option is removed. That
means DirEntry.metadata() is never a cached value. Instead you want to read
metadata you should do it in the process_entries callback and store whatever
values you need as client_state. See this [benchmark] as an example.
<csr-read-only-do-not-edit/>
<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
cf9d248)6410407)85dff6e)0dfefa9)797f76f)ff8f491)898ff91)b330914)0f8b1c8)f9a144b)11fa0bc)3b90d8e)850954f)45529ef)66d46ac)7ee0cb3)17ddb91)4e4218f)1074566)064ee60)8683484)0ef72d1)fd9af20)c10dbaa)99678be)29b6b1e)
</details>3 commits contributed to the release.
<csr-read-only-do-not-edit/>
<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
8 commits contributed to the release over the course of 9 calendar days.
<csr-read-only-do-not-edit/>
<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
09ebe45)f47607d)b8e9aea)29c035e)331a896)c1bffdd)e0d3e3e)133d168)
</details>4 commits contributed to the release.
<csr-read-only-do-not-edit/>
<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
37 commits contributed to the release over the course of 13 calendar days.
<csr-read-only-do-not-edit/>
<csr-read-only-do-not-edit/>
<details><summary>view details</summary>
ad9404e)8fd5ae1)f432c4f)347bc92)0bb13ee)ff5fba1)0b23848)0ee3b3c)b3ac65c)301e2a5)ab07f7a)66828d7)58c7707)566da8f)ade977d)738d9f6)1f42a3e)5fc2859)5cc9db7)2e58da7)5cd8e04)c6f8385)6372a1a)32a527d)d3899aa)32d6b86)5b4745e)3ecc5b6)5d9ee63)4ced894)ebf9789)36b2d2f)6b0df4a)115a5fe)ba02dda)b63f810)8f64d5a)
</details>Nothing published for this version
Nothing published for this version
One column per quarter.