Closed muzarski closed 2 months ago
cargo semver-checks
detected some API incompatibilities in this PR.
Checked commit: ce0fdca7f917598efd5b60e0c96316c2648398a4
See the following report for details:
The problems you mentioned sound like serious bugs. Are they present in 0.14 or only on main?
The problems you mentioned sound like serious bugs. Are they present in 0.14 or only on main?
From what I see, they are directly caused by PRs merged in recent days.
The problems you mentioned sound like serious bugs. Are they present in 0.14 or only on main?
It was merged yesterday: https://github.com/scylladb/scylla-rust-driver/pull/1067, so only main
is affected
v1.1: Deduplicated the logic shared between ClusterWorker
and PoolRefiller
Ref: https://github.com/scylladb/scylla-rust-driver/issues/519
Motivation
The error refactor that I'm currently working on, surprisingly changed driver's logic in some cases where the decision is made based on the error returned from query execution. There are three places where driver's logic was silently altered:
During error refactor I changed all places where
QueryError::IoError
was constructed. Now all of these IoErrors are represented asBrokenConnectionError
andConnectionPoolError
types (and their variants). These types represent an error which implies that a corresponding connection/node is broken/unreachable.Changes
[Query/NewSession]Error::IoError
variants altogether. They are no longer constructed anywhere - they were replaced byConnectionPoolError
andBrokenConnection
variants during previous refactors.Pre-review checklist
./docs/source/
.~Fixes:
annotations to PR description.