The Blog

The maintenance release includes a regression test for observed wrapped arrays and changes the Linux runtime build to avoid libatomic.

The 3.9 release candidate changes how optional chaining combines with a non-null assertion. An assertion about types no longer inserts an unintended runtime boundary in the chain.

April's development snapshot updates core modules while giving maintainers a concrete interpreter for testing comparison chains and Unicode name handling.

PgBouncer 1.13 adds a Linux TCP timeout, changes permitted TLS versions and reports readiness to systemd. Each change controls a different part of connection management.

An istiod disruption budget selected a label absent from its pods. Istio 1.5.2 repairs that mismatch and several installation paths introduced around the consolidated control plane.

Cassandra 4.0-alpha4 bounds repair-history retention and adds repair diagnostics. Its preview also exposes the consistency choices behind foreground read repair.

Cargo supplies executable paths when building integration tests and benchmarks. Smaller inference and macro changes also remove cases where valid intent needed extra work.

Datasette 0.40 removes the Now v1 publisher and directs Now v2 users to a plugin. It also accepts YAML metadata and fixes configuration handling during publishing.

The second 20.1 release candidate prevents mixed-version clusters from accepting time precision that older nodes ignore. The repair illustrates why new binaries and a finalized cluster version are different states.

Cluster API v1alpha3 adds declarative kubeadm control plane management and bounded worker health remediation. Its APIs distinguish control plane sequencing from infrastructure-managed worker groups.

Markhor updates the system base while documenting the state and permission transitions that a declarative rebuild cannot infer on its own.

The final Python 2 release includes a repair for control characters in HTTP hostnames. Its contents explain why an available release and a supported runtime are different claims.

The 2019 survey finds broad module adoption alongside persistent tooling problems. Its revised methodology also changes how several apparent year-to-year trends should be read.

Co-groups let several grouped streams update one aggregate with separate input functions. TLS 1.3 support arrives alongside a distinct change making TLS 1.2 the default.

P0533R6 proposes a bounded extension of constant evaluation into cmath and cstdlib. Runtime environment dependencies explain why the proposal selects some functions and excludes others.

rqlite 5.2.0 fixes a zero snapshot-check interval that allowed unbounded Raft log growth. The check interval and entry threshold govern different parts of the maintenance process.

N2510 removes the requirement to name every parameter in a C function definition. Callback implementations can retain the required signature without introducing an unused identifier.

Opening the Compose specification creates a place to define what an application file means across implementations, including where platform support can differ.

The April announcement confirms that Ruby 2.4 support ended on March 31, after a final security release, and distinguishes the remaining branches by maintenance policy.

Sentinel's redesigned Terraform imports are generally available. Their relationship to Terraform's JSON representation explains the policy decisions that require more than a final value.

CVE-2020-10933 comes from changing a reusable string buffer before a nonblocking read succeeds. The fix separates capacity from the visible string length.

Redis 6.0 RC3 fixes a critical connection-management crash and introduces ordered timeout tracking for blocking commands. The timeout implementation separates request waiting from idle-connection handling.

Beta 4 fixes writes racing with column backfills and restricts node-local uploads to administrators. The changes clarify how online schema work and operational file access cross different boundaries.

rqlite 5.1.1 corrects how freshness interacts with consistency levels. The implementation now applies its leader-contact threshold only to reads using level none.

ts-expect-error reports an unused directive when the following line stops producing a diagnostic. That gives negative type tests and temporary workarounds an explicit expiry condition.

Topology Manager reaches beta with a node admission policy for NUMA alignment. Kubernetes 1.18 also adds alpha debugging through ephemeral containers and improves the tools used to inspect resource changes.

Datasette 0.39 adds base_url for deployments under a proxy path and table metadata for choosing the initial sort order.

Clang 10 implements concepts under -std=c++2a. Its experimental caching switch exposes the interaction between constraint semantics and compile-time cost.

Swift 5.2 improves error diagnosis inside the type checker and adds more direct ways to express property projections in APIs.

The new PostgreSQL APT archive makes historical packages available through repository indexes. Package selection still depends on APT policy and the distribution being reconstructed.