# TODO
* Features
* 可以用过程宏实现：impl Disposable for Shared\<Mutable\<Context\>\>
* Async 转换 + Single 抽象（[最后验证操作符是不是全](https://reactivex.io/documentation/single.html)）
* Future 转 Single，可以删掉 Future 转 Observable
* Single 转 Future，可以删掉 Observable 转 Future（应该没有这个）
* NotBigDeal
* 用 test_channel 代替 test 里的PublishSubject和create，just
* Infallible 替换为Never
* 去掉test_without_convenient_api
# Check List
1. 注释
2. 文档 Example
3. 合理使用（减少不必要）
1. pub，move
2. Clone，Send, Sync, 'static
3. lock/read 等锁。 检查是否需要用Context来用一个锁？不然可能有并发问题
4. 锁的使用
1. 尽量使用safe_lock宏
2. 检查锁是否及时释放
1. 在 on_next 时释放context锁
2. 在 on_termination 时，释放context锁和observer锁
3. 检查锁是否应该更大范围，即一个“事务”内，只锁一次。
4. 单测
1. Cases:
1. test_completed
2. test_error
3. test_unsubscribe
4. test_ref
5. test_mut_ref
6. test_async
7. test_subscribe_by_different_observer
8. test_unsub_on_next_by_take
9. Non-creating observable
1. test_multiple_operation
2. test_without_convenient_api
10. Revertible observable
1. test_revert_completed
2. test_revert_error
11. Hot Observable (Subjects, RefCount)
1. test_complete_on_next
2. test_error_on_next
3. test_unsub_on_next
4. test_sub_on_next
5. test_next_on_next
6. test_unsub_on_completed
7. test_sub_on_completed
8. test_unsub_on_error
9. test_sub_on_error
12. Using lock(lock_mut|safe_lock.*!) 有的lock没 test_complete_after_next
1. test_immediate_next
2. test_immediate_completed
3. test_immediate_error
13. Using schedule(: Scheduler|::from_stream\\() 有的scheduler没有test_immediate_next （如 from future）
1. TESTS IN USING LOCK
2. test_complete_after_next
3. test_error_after_next
4. test_unsub_after_next
5. test_unsub_after_completed
6. test_unsub_after_error
7. test_order_with_continuous_next
8. test_no_delay (if appliable)
9. test_undisposed_scheduler (search this to find all module using scheduler)
10. test_scheduler_should_be_disposed_after_completed
11. test_scheduler_should_be_disposed_after_error
12. test_scheduler_should_be_disposed_after_unsub
14. Compiling checking
1. test_lifetime
2. test_fn
3. test_clone
4. test_type_inference_with_subscribe
5. test_type_inference_without_subscribe
2. 覆盖率
# Commend
* cargo fmt
* cargo clippy --fix
* cargo hack check --each-feature --no-dev-deps --exclude-all-features
* reset; time RUST_BACKTRACE=1 cargo nextest run --no-fail-fast --failure-output final --status-level=fail --no-default-features --features tokio-scheduler
* fd7a1e528b7c241bcc21617c8c243d421fc6252f
* RUST_BACKTRACE=1 cargo nextest run --no-fail-fast --failure-output final 119.88s user 20.74s system 289% cpu 48.605 total
* 5b827281fe3626ce45fd0771e2c1ebadf05f2166
* RUST_BACKTRACE=1 cargo nextest run --no-fail-fast --failure-output final 183.86s user 28.95s system 372% cpu 57.055 total
* 6dd8a84347006e2525c49ef3eaf5bb281535880f
* RUST_BACKTRACE=1 cargo nextest run --no-fail-fast --failure-output final 200.73s user 29.72s system 340% cpu 1:07.67 total
* reset; time RUST_BACKTRACE=1 cargo hack --each-feature --exclude-all-features --exclude-no-default-features --exclude-features default,multi-threaded,single-threaded nextest run --no-fail-fast --failure-output final --status-level=fail
* cargo test --doc --features tokio-scheduler
* cargo tarpaulin --out Html --features tokio-scheduler
* cargo tarpaulin --out Html -- --test utils::unique_key_store
* cargo doc --open
* cargo expand