Release 3.10.0: changelog entries for the PRs merged after the third splice

[erikdarlingdata]

Changelog entries for the 3.10.0 PRs merged to dev after the third splice (#5576). No issue. No dev to main PR, no tag, no release. The ## [3.10.0] heading and date are untouched.

What changes

Fourteen entries, text unchanged from the changelog buffer. Each full entry is appended at the end of its section in docs/changelog/3.10.md. Its title-only index line goes at the end of the same section in CHANGELOG.md, with a link definition for each reference. All three sections (Added, Changed, Fixed) already existed under 3.10.0, so none was added. tools/changelog/archive-census.txt is regenerated (3.10.0: 276 to 290 entries).

• Added: Column filters list a column's values to tick, like Excel, and stay set across refreshes and restarts #5577, Time range picker: more presets, calendar periods and typed ranges #5570
• Changed: Query Store interval tables are partitioned by day #5573, Perfmon data is compressed per counter, so a one-counter trend reads only that counter's data (#5574) #5579, The first retention drain gives way to collection, so it no longer makes slots skip (#5592) #5595
• Fixed: The module map refresh no longer holds back the collectors at start (#5578) #5580, The module map refresh groups procedure_stats before its DISTINCT ON (#5583) #5584, After a big retention drain, warm each server's Query Store floor read in the background (#5581) #5585, Web charts draw at the panel's real width, so axis text is not stretched (#5586) #5590, Query Store interval tables stop cancelling their own autovacuum (#5589, #5594) #5593, Test processes no longer pay 0.3 s per WPF STA thread at exit, which failed green Darling runs (#5596) #5598, Collection Falling Behind no longer fires in the first hour after every service start (#5597) #5599, A hung Darling or Lite test shard names its test and still uploads its logs (#5587) #5591, Query Store rollup: the build wait counts only writers of the wide table, and the migration lock log lists every lock holder (#5582) #5610
• Query Store rollup: the build wait counts only writers of the wide table, and the migration lock log lists every lock holder (#5582) #5610 keeps both references, ([#5588], [#5610]), and both have a link definition.

Only the three allowed files changed. The script that inserted the entries was run from scratch space and is not committed.

Accounting

Fresh list of PRs merged to dev since 2026-10-08T17:15Z: 17 PRs.

PR	Record
#5570, #5577	Added entry
#5573, #5579, #5595	Changed entry
#5580, #5584, #5585, #5590, #5591, #5593, #5598, #5599	Fixed entry
#5610	Fixed entry, which names both #5588 and #5610
#5588	No entry of its own, by design: the entry is #5610's
#5604, #5609	No entry, by design: test-only

No PR merged after the list was written without a buffer line.

Entry-line proof

• Buffer, (section, house-style line) for the 14 PRs: 14 lines, SHA-256 of the sorted lines 495da3797ce267fb5716d3b384c050dbc2aa83bbeb09de5cdc9d197ec94b9db2.
• Lines docs/changelog/3.10.md gained under 3.10.0 (Fixed 9, Changed 3, Added 2): 14 lines, the same SHA-256. Nothing was lost.
• Index lines CHANGELOG.md gained: 14 title-only lines (SHA-256 fa64c43993e863d9734d8535a1e29b2026a2fa1843c42cbab131b259f87655f3), as split writes them. It also gained 15 link definitions.
• Byte census after the write: docs/changelog/3.10.md 623 CRLF, 0 bare CR, 0 bare LF. CHANGELOG.md 5189 CRLF, 0 bare CR, 0 bare LF (the blobs are LF, and .gitattributes makes the working tree CRLF).

Archive script

python tools/changelog/changelog_archive.py split wrote 12 files and changed nothing beyond the three above. Last lines of each command:

• split: wrote 12 files
• census: wrote tools/changelog/archive-census.txt
• verify: PASS 3.9.0: 760 entries, prose sha256 3b0001be58413502 matches tools/changelog/archive-census.txt
• roundtrip: PROSE INTACT every version 3653e8d63a08 held is byte-identical in prose, and every link definition still resolves

Test plan

• Build of Darling.Tests (Debug): 0 Warning(s), 0 Error(s).
• ChangelogIndexAndArchiveTests: Total 5, Failed 0.
• ProductVersionDeclarationTests: Total 26, Failed 0.
• ReleaseStatementGateWorkflowTests: Total 4, Failed 0.
• Every *Census* class: Total 499, Failed 0 (1 skipped).
• -trait Stage=Guard -failSkips: Total 1467, Failed 0.
• Full suite: not run on the laptop. CI decides.

For the seat

• The seat merges this. No merge-train label was added.