vincenzopalazzo
/
arti
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
arti/doc/semver_status.md

52 lines
1.2 KiB
Markdown
Raw Normal View History

Add a file to track per-crate semver status.
2021-11-11 15:35:38 +00:00
# Semver tracking
This is a helpful file that we use for checking which crates will have
breaking or nonbreaking API changes in the next release of Arti.
For each crate, please write "BREAKING" if there is an API change that counts
as breaking in semver, and "MODIFIED" if there is a backward-compatible API
change.
You can change from MODIFIED to BREAKING, but never from BREAKING to
MODIFIED.
You don't need to list details; this isn't the changelog.
Don't document other changes in this file.
We can delete older sections here after we bump the releases.
Clean out semver_status.md
2022-04-01 16:18:35 +01:00
## Since Arti 0.2.0
Add a file to track per-crate semver status.
2021-11-11 15:35:38 +00:00
Add RetrySchedule::reset() Previously the code would do stuff like ``` schedule = RetrySchedule::new(INITIAL_DELAY); ``` which is needlessly verbose, since the schedule already keeps track of its initial delay.
2022-04-04 15:27:51 +01:00
### tor-basic-util
MODIFIED: Added `reset()` method to RetrySchedule.
chanmgr: bubble ClockSkew up through the Error object. Fortunately, we don't need a separate type here: authenticated clock skew can only come attached to a `tor_proto::Error`. We also remove skew from `tor_proto::Error::HandshakeCertsExpired`, since it would now be redundant.
2022-04-07 15:01:19 +01:00
### tor-chanmgr
BREAKING: Added members to `Error::Proto`
ChanMgr: Return provenance information from get_or_launch We need this since we want to report certain conditions only when they happen on a new channel, not if we observe them on a preexisting channel.
2022-04-07 15:43:04 +01:00
BREAKING: Added `ChanProvenance` to `ChanMgr::get_or_launch`.
chanmgr: bubble ClockSkew up through the Error object. Fortunately, we don't need a separate type here: authenticated clock skew can only come attached to a `tor_proto::Error`. We also remove skew from `tor_proto::Error::HandshakeCertsExpired`, since it would now be redundant.
2022-04-07 15:01:19 +01:00
circmgr: Avoid a race condition in circuit usage restriction We were treating restrict_mut() failures as internal errors, and using internal errors to represent them. But in fact, these failures are entirely possible based on timing. Here's how it happens: * Two different circuit requests arrive at the same time, and both notice a pending circuit that they could use. * The pending circuit completes; both pending requests are notified. * The first request calls restrict_mut(), and restricts the request in such a way that the second couldn't use it. * The second request calls restrict_mut(), and gets a failure. Because of this issue, we treat these errors as transient failures and just wait for another circuit. Closes #427. (This is not a breaking API change, since `AbstractSpec` is a crate-private trait.)
2022-03-31 19:51:02 +01:00
### tor-circmgr
socksproto: remove some unused accessors.
2022-04-02 17:01:20 +01:00
circmgr: Avoid a race condition in circuit usage restriction We were treating restrict_mut() failures as internal errors, and using internal errors to represent them. But in fact, these failures are entirely possible based on timing. Here's how it happens: * Two different circuit requests arrive at the same time, and both notice a pending circuit that they could use. * The pending circuit completes; both pending requests are notified. * The first request calls restrict_mut(), and restricts the request in such a way that the second couldn't use it. * The second request calls restrict_mut(), and gets a failure. Because of this issue, we treat these errors as transient failures and just wait for another circuit. Closes #427. (This is not a breaking API change, since `AbstractSpec` is a crate-private trait.)
2022-03-31 19:51:02 +01:00
MODIFIED: Added a new variant in tor_circmgr::Error.
rtmock: add the ability to make a connection time out.
2022-04-02 20:48:41 +01:00
Implement a better clock skew estimator. This time, our estimator discards outliers, takes the mean of what's left, and uses the standard deviation to try to figure out how seriously to take our report of skew/not-skew. These estimates are still not actually used.
2022-04-07 21:03:25 +01:00
### tor-guardmgr
MODIFIED: New functions to get estimated clock skew.
MODIFIED: New functions to report observed clock skew.
Channel: Expose our view of whether the clock is skewed, and the age of a channel. At first I wanted to have this information not be a part of channels at all, but it is a fairly tiny amount of data, and the alternatives are pretty crufty.
2022-04-06 20:29:24 +01:00
### tor-proto
MODIFIED: New accessors in tor_proto::Channel.
chanmgr: bubble ClockSkew up through the Error object. Fortunately, we don't need a separate type here: authenticated clock skew can only come attached to a `tor_proto::Error`. We also remove skew from `tor_proto::Error::HandshakeCertsExpired`, since it would now be redundant.
2022-04-07 15:01:19 +01:00
BREAKING: Removed clock skew from Error::HandshakeCertsExpired.
Implement a better clock skew estimator. This time, our estimator discards outliers, takes the mean of what's left, and uses the standard deviation to try to figure out how seriously to take our report of skew/not-skew. These estimates are still not actually used.
2022-04-07 21:03:25 +01:00
MODIFIED: New functions on ClockSkew.
Channel: Expose our view of whether the clock is skewed, and the age of a channel. At first I wanted to have this information not be a part of channels at all, but it is a fairly tiny amount of data, and the alternatives are pretty crufty.
2022-04-06 20:29:24 +01:00
rtmock: add the ability to make a connection time out.
2022-04-02 20:48:41 +01:00
### tor-rtmock
MODIFIED: Added add_blackhole to MockNetwork.
Merge branch 'main' into 'retriable' # Conflicts: # doc/semver_status.md
2022-04-05 15:48:51 +01:00
### tor-socksproto
BREAKING: Removed some unused accessors.