All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog and this project adheres to Semantic Versioning.
- Replace dependency on threadpool crate with a custom solution built on the standard library only, and only using scoped threads -> fixes memory leaks observed when running under valgrind
- up MSRV to 1.63 for scoped threads
0.7.3 - 2024-05-10
- Default to single-threaded tests for WebAssembly (thanks @alexcrichton) in #41
0.7.2 - 2024-04-09
- Fix
Conclusion::exit_code
(logic was inverted in 0.7.1)
0.7.1 - 2024-04-09
- Add
Conclusion::exit_code
and note about destructors/cleanup to docs ofexit
andexit_if_failed
e938e537e
0.7.0 - 2024-01-14
- Also check
kind
when filtering tests (thanks @sunshowers) in #30- This is potentially breaking as additional or fewer tests might be executed in some situations.
- Add JSON format output (thanks @PaulWagener and @t-moe) in #35
- Add no-op flags to add CLI compatibility for IntellJ Rust (thanks @Dinnerbone and @t-moe) #28 /
70cdc55
0.6.1 - 2022-11-05
- Actually spawn as many threads as specified by
--test-threads
(thanks @hgzimmerman) in #32. - Fix & improve docs
- Fix badge in README
- Deemphasize MSRV by removing check and note from README in #24.
0.6.0 - 2022-11-05
- Breaking: Updated
clap
to version 4 (thanks @msrd0) - Breaking: Bump MSRV to 1.60 (due to the clap update)
- Breaking: Remove
FromStr
impls forargs::{ColorSetting, FormatSetting}
(useclap::ValueEnum
instead).
0.5.2 - 2022-08-14
- Re-add
--nocapture
as a noop argument #18 (thanks @sunshowers)
- Link in documentation
0.5.1 - 2022-08-13
Trial::{name, kind, has_ignored_flag, is_test, is_bench}
getters
0.5.0 - 2022-08-13
Most parts of this library have been rewritten and the API has changed a lot. You might be better of just reading the new docs instead of this change log. I do think the new API is better in many regards. Apart from an improved API, changes that motivated the rewrite are marked with ⭐.
- Breaking: bump MSRV to 1.58
- Breaking: Rename
Test
toTrial
- Breaking: Rename
run_tests
torun
- ⭐ Breaking: Make every
Trial
have a runner function instead ofdata
+ a global runner function. Thus, the third parameter ofrun
is no more. I think this model is more intuitive. - Breaking: Add
Trial::{test, bench}
constructor functions, use builder pattern, and make fields private. - Breaking: rename
Args::num_threads
totest_threads
- Breaking: make fields of
Conclusion
public and remove getter methods - Breaking: remove
RunnerEvent
. This should not have been public. - ⭐ Tests are now run in main thread when
--test-threads=1
is specified - ⭐ Reduce number of indirect dependencies considerably
- Fix
rust-version
field inCargo.toml
(thanks @hellow554) - Fix
--ignored
behavior - Fix some CLI error messages
- ⭐Panics in test runners are caught and treated as failure
- ⭐ Lots of integration tests (should make any future development of this library way easier)
- Add
must_use
message forConclusion
- Print total execution time at the end of the run
- Allow benchmarks to run in test mode
--include-ignored
- Breaking: remove unsupported CLI options. They were ignored anyway, but the CLI would accept them.
0.4.1 - 2022-06-07
- Add
rust = "1.56"
toCargo.toml
, stating the existing MSRV. - Update
crossbeam-channel
to deduplicate some indirect dependencies.
0.4.0 - 2022-05-13
- Breaking: Update to Rust 2021, bumping MSRV to 1.56
- Fix
--list --ignored
behavior
0.3.0 - 2020-06-28
- Add support for running tests in parallel #4
- Add
Arguments::from_iter
#5
0.2.0 - 2019-10-02
- Upgrade dependencies #3
- Flush stdout after printing test name 4a36b3318b69df233b0db7d1af3caf276e6bb070
- Fix overflow bug when calculating number of passed tests 264fe6f8a986ab0c02f4a85e64e42ee17596923c
- Everything.