PostgreSQL: Clearing the Deck for Version Twenty
Postgres eighteen development dropped support for pre-version-ten servers across pg-dump, pg-upgrade, and psql, while separate fixes addressed misleading wraparound advice, a repack bug with generated columns, and protocol tracing errors.
Duration: PT2M25S
Episode overview
This episode is a short developer briefing from PostgreSQL.
It explains recent repository work in plain language.
- Show: PostgreSQL
- Published: 2026-07-03T14:13:48Z
- Audio duration: PT2M25S
Transcript excerpt
This excerpt keeps the crawler page concise. Listen to the episode or use the RSS feed for the full update.
Good day. It's July third, twenty twenty-six, and this is PostgreSQL.
The headline today is housekeeping with teeth: the project is formally dropping support for servers older than version ten across its core tools, ahead of the version twenty release.
Nathan Bossart landed three related commits pulling pre-ten support out of pg-dump and pg-dumpall, pg-upgrade, and psql. Commit three-a-zero-a-three-zero-eight, one-four-d-eight-four-one-eight, and eight-three-one-b-e-c-four all follow the same pattern: the project maintains a ten-version support window, and with…
Second theme: reducing misleading guidance for operators. Fujii Masao's commit zero-eight-four-seven-three-four removes replication slot advice from MultiXact wraparound hints. That advice made sense for transaction ID wraparound but not MultiXact, where dropping slots doesn't actually resolve the underlying…
Fujii also authored the biggest feature of the batch: a new log_statement_max_length setting, in commit c-eight-b-d-eight-three-eight, letting administrators cap how much statement text gets written to logs. It's byte-based but respects character boundaries, and it leaves error-statement logging untouched for now.
A few…