Skip to content

Add the missing [Unreleased] CHANGELOG entries for eleven merged PRs - #4438

Merged
erikdarlingdata merged 1 commit into
devfrom
docs/changelog-unreleased-catch-up
Sep 26, 2026
Merged

erikdarlingdata merged 1 commit into
devfrom
docs/changelog-unreleased-catch-up

Conversation

@erikdarlingdata

@erikdarlingdata erikdarlingdata commented Sep 26, 2026 •

Copy link
Copy Markdown
Owner

Adds the [Unreleased] entries that eleven PRs merged since v3.8.0 were missing, or records why a PR needs none.

What changes

PR decision section one-line reason from the diff
#3588 entry Fixed compression-stuck self-alert paged on a healthy job's own run instant; now confirms a second read before reporting
#3886 entry Fixed job_history collector's dedup watermark outlived an msdb reseed's identity reset, silently losing job history forever
#3889 entry Fixed collection health banded HEALTHY on a collector that produced then went to zero rows for weeks
#3900 entry Fixed per-server collection-health memo keyed on window start time, never hit in production, and grew unbounded
#3911 entry Fixed fleet card's collection-health read decompressed a week of raw log rows on every call; now served from an hourly rollup
#4016 entry Fixed plan capture silently stopped matching under a custom log_line_prefix with more than one field before the pid
#4032 entry Fixed trace-flag reads compared the wrong timestamps and hid every enabled flag after an ordinary run
#4042 entry Fixed deadlock/plan-capture log patterns were wider than the assembler's own rule; narrowed so a crafted statement is never offered as a candidate report
#4157 entry Changed llms.txt and CITATION.cff had several stale facts (collector counts, supported platforms, alert channels, download count, vendor name)
#4332 None — adds RawChunkIntervalPlanner, a pure planner with no database wiring yet ("No database wiring yet - wiring this into the TimescaleDB ensure path... is a separate change"); doc-comment rewording only otherwise
#4398 None — test-only fix to how a DuckDB memory-trim test measures memory; no product code touched

Test plan

  • ChangelogIndexAndArchiveTests: 5/5 passed on 9e9ce79 merged with dev 13329d1 (the real Darling.Tests.dll, run in-process). CI takes the docs fast path for a CHANGELOG-only change, so it doesn't run this class.
  • All eight test classes that read CHANGELOG.md also pass on that merge: 405 total, 0 failed, 23 skipped (live-PostgreSQL tests only).
  • DocCommentHygieneTests (Darling.Tests, in-process on macOS): 77/77 passed. (No doc comments, members, or log text were touched by this PR, so this is a confirming run, not a pin against new risk.)
  • python3 .github/scripts/ci_description_symbol_drift.py --help: this script checks a PR's own description against its own diff; it takes no local CHANGELOG input, so it does not apply to this change and was not run against these entries.
  • Lite.Tests/CrossAppGuardCiGateTests.cs: grepped for CHANGELOG; the only hit is a doc-comment sentence about test fixtures having "no path to add" — it does not assert anything about CHANGELOG.md's content, so nothing in that class exercises this change.
  • dotnet build Darling/Darling.Tests/Darling.Tests.csproj -c Release -p:EnableWindowsTargeting=true: 0 errors, 25 pre-existing warnings unrelated to this change.
  • No other file was touched; git diff --stat shows only CHANGELOG.md, +18/-0.

CHANGELOG entry

SECTION: None — this PR is the CHANGELOG change itself.

Adds Fixed entries for #3588, #3886, #3889, #3900, #3911, #4016, #4032 and
#4042, and a Changed entry for #4157. #4332 (an unwired chunk-interval
planner) and #4398 (a test-only measurement fix) get no entry: neither
changes anything a user of the shipped product can observe.
@erikdarlingdata
erikdarlingdata marked this pull request as ready for review September 26, 2026 17:36
@erikdarlingdata
erikdarlingdata merged commit 2c34fc0 into dev Sep 26, 2026
17 of 18 checks passed
@erikdarlingdata
erikdarlingdata deleted the docs/changelog-unreleased-catch-up branch September 26, 2026 17:36
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant