feat: tokio upgrade #729
Merged
feat: tokio upgrade #729
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
koivunej
reviewed
Aug 2, 2021
koivunej
reviewed
Aug 2, 2021
How come this didn't manifest itself in test-md build...? Surely it does a lot of filters. Did you add unconditional boxing to the filters and I missed it while doing the commit-per-commit? |
I did try inserting
I'm not sure about |
|
Force pushed to include |
Merged
koivunej
reviewed
Aug 3, 2021
|
Rebase'd master changes (CI speed improvements) |
|
Commit e72bd5c introduces a work-around fix for the warp stack-overflow bug (mentioned in PR description). Further investigation indicates that no other stack-overflow issues remain (from this bug) -- used test coverage to check all instances of filters are covered or manually inspected if missed. |
koivunej
approved these changes
Aug 11, 2021
Signed-off-by: Mirko von Leipzig <[email protected]>
Required manually implementing Serialization for `Secret<String>` as this is purposefully disabled by default. Signed-off-by: Mirko von Leipzig <[email protected]>
Main API changes are `bytes()` to `chunks()` and similar renamings. Signed-off-by: Mirko von Leipzig <[email protected]>
Does not compile, requires changes to many other dependencies as well. Main API change is `delay_for` to `sleep`. Signed-off-by: Mirko von Leipzig <[email protected]>
`DelayQueue` was moved from `tokio` to `tokio-util` in preparation for the 1.0 release of `tokio`. Signed-off-by: Mirko von Leipzig <[email protected]>
This fixes the missing `Stream` adapters for `tokio` related items. For example, `tokio::sync::Receiver<T>` used to implement `FutureExt` trait, but it no longer does. Instead we now wrap it using `tokio-stream` to provide the same extension. Signed-off-by: Mirko von Leipzig <[email protected]>
Signed-off-by: Mirko von Leipzig <[email protected]>
Signed-off-by: Mirko von Leipzig <[email protected]>
Signed-off-by: Mirko von Leipzig <[email protected]>
Signed-off-by: Mirko von Leipzig <[email protected]>
Signed-off-by: Mirko von Leipzig <[email protected]>
Note that version "0.3.15" was yanked. Signed-off-by: Mirko von Leipzig <[email protected]>
Signed-off-by: Mirko von Leipzig <[email protected]>
Signed-off-by: Mirko von Leipzig <[email protected]>
Signed-off-by: Mirko von Leipzig <[email protected]>
Also includes removing the base tungstenite dependency. Since tokio-tungstenite re-exports it. Signed-off-by: Mirko von Leipzig <[email protected]>
Previous version was yanked. Changes are - `Domain::ipv4()` -> `Domain::IPV4` - `Type::stream()` -> `Type::STREAM` - `socket.into_tcp_listener()` -> `TcpListener::from(socket)` Signed-off-by: Mirko von Leipzig <[email protected]>
Signed-off-by: Mirko von Leipzig <[email protected]>
futures-retry API changes: - `#attempts` is now tracked by the library - retry output changed: `Result<output, err>` to `Result<(output, attempts), (err, attempts)>` Signed-off-by: Mirko von Leipzig <[email protected]>
Maximum retries were being passed in to the RequestErrorHandler as is. However in RequestErrorHandler it is defined as maximum attempts. i.e. `N > max` versus `N == max` In the future-retries upgrade the check for this was corrected to match the attempts naming, but this broke the actual usage which assumed it was retries. Signed-off-by: Mirko von Leipzig <[email protected]>
Signed-off-by: Mirko von Leipzig <[email protected]>
Fix for stack-overflow in debug mode, caused by too many warp filters. Issue link: seanmonstar/warp#811 Signed-off-by: Mirko von Leipzig <[email protected]>
Was a tokio 0.2 dependency, no longer needed. Signed-off-by: Mirko von Leipzig <[email protected]>
The error handler is rarely / never configured for tests, and this error was therefore not visible. Signed-off-by: Mirko von Leipzig <[email protected]>
Part of upgrading tokio from v0.2 to v1.*. Signed-off-by: Mirko von Leipzig <[email protected]>
Patched for versions >=0.14.10, and we now use 0.14.11. Signed-off-by: Mirko von Leipzig <[email protected]>
Patched in versions >=0.6.3, we use 0.6.4. Signed-off-by: Mirko von Leipzig <[email protected]>
This now matches the version used by our dependencies, so that we only depend on one version overall. Signed-off-by: Mirko von Leipzig <[email protected]>
This matches the version used by warp, and means we only depend on a single version. The downgrade doesn't remove any functionality we rely on. Signed-off-by: Mirko von Leipzig <[email protected]>
|
Allrighty, this is ready to merge. This a big step forwards, thanks a lot! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Quite a large and difficult refactor. There are some issues remaining which require feedback.
todo!I left for format to strings where I don't know what info is valuable.Test status remained the same for me before and after. Tests can all pass; but the
ilp-nodetests seem plagued by racey start-up conditions (?) and sporadically fail -- this behaviour was the same before the refactor.--edit--
Added a fix for (1) -- tests now pass in debug mode. We will want to ensure that other filter handlers are covered by tests, but that should be a new issue.