<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Release-Engineering on Kuldeep Pisda</title><link>https://kdpisda.in/tag/release-engineering/</link><description>Recent content in Release-Engineering on Kuldeep Pisda</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Sun, 20 Sep 2026 09:00:00 +0530</lastBuildDate><atom:link href="https://kdpisda.in/tag/release-engineering/index.xml" rel="self" type="application/rss+xml"/><item><title>PostgreSQL's Graph Queries Took Two Years to Build. An AI Bug Tally Killed Them in Nine Days</title><link>https://kdpisda.in/postgresql-19-sql-pgq-revert-scary-patch-contest/</link><pubDate>Sun, 20 Sep 2026 09:00:00 +0530</pubDate><guid>https://kdpisda.in/postgresql-19-sql-pgq-revert-scary-patch-contest/</guid><description>&lt;p&gt;On September 7, PostgreSQL&amp;rsquo;s release team reverted SQL/PGQ, the built-in property-graph query feature, out of the PostgreSQL 19 branch. Forty-seven commits gone: the feature itself and every post-commit fix layered on top of it since March. It had been in development since February 2024. It shipped in beta. It is not shipping in 19.&lt;/p&gt;
&lt;p&gt;The proximate cause wasn&amp;rsquo;t a single catastrophic bug report. It was a pgsql-hackers thread from August 25 titled &amp;ldquo;scary patch contest,&amp;rdquo; in which Robert Haas, one of PostgreSQL&amp;rsquo;s longest-serving committers, wrote that he&amp;rsquo;d asked Claude to rank the release&amp;rsquo;s patches by how many bugs had been fixed in them since feature freeze. SQL/PGQ was one of six patches the exercise flagged. Nine days after Andres Freund replied to that thread with his own independent findings, SQL/PGQ was gone.&lt;/p&gt;</description></item></channel></rss>