chore: bump pg_ducklake and DuckDB to v1.5.1 (#60)

## Summary

- Bump `DUCKLAKE_COMMIT` to `d5dd630` (pg_ducklake latest main, includes
DuckDB v1.5.1)
- Update `duckdb` Rust crate from `1.10500.0` to `1.10501.0`
- Update `DUCKDB_VERSION` from `v1.5.0` to `v1.5.1` in Docker configs

## Test plan

- [x] `cargo update -p duckdb` — lock file updated cleanly
- [x] `make build` — compiles against libduckdb v1.5.1
- [x] `make check-regression` — all 41 regression tests pass

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: YuweiXiao

.claude/skills/bump-pgducklake/SKILL.md

@@ -0,0 +1,128 @@
+---
+name: bump-pgducklake
+description: Bump pg_ducklake (and optionally DuckDB) to a new version. Updates all pinned commits and version constants across the repo.
+user_invocable: true
+allowed_tools: Bash, Read, Edit, Glob, Grep, WebFetch, AskUserQuestion
+---
+
+# /bump-pgducklake — Bump pg_ducklake / DuckDB Version
+
+Bump the pg_ducklake commit pin (which transitively bumps DuckDB, pg_duckdb, and the ducklake extension) across all files in the repo.
+
+**IMPORTANT**: All paths below are relative to the **repo root** (the directory containing `Cargo.toml`, `Makefile`, and `duckpipe-pg/`). Determine the repo root at the start — if working inside a worktree, use the worktree root.
+
+---
+
+## Step 0 — Determine the Target
+
+Ask the user or infer from context:
+- **pg_ducklake commit**: the target commit hash on `https://github.com/relytcloud/pg_ducklake.git` (default: latest `main`)
+- **DuckDB version**: the DuckDB version that commit includes (check the pg_ducklake commit log for version bump commits)
+
+If the user just says "bump to latest", fetch the latest commit:
+```bash
+git ls-remote https://github.com/relytcloud/pg_ducklake.git refs/heads/main
+```
+
+To find the DuckDB version, check the pg_ducklake repo's recent commits or `third_party/pg_duckdb/third_party/duckdb` submodule.
+
+---
+
+## Step 1 — Determine the duckdb-rs Crate Version
+
+The `duckdb` Rust crate uses the version scheme `1.{10000*major + 100*minor + patch}.0`. For example:
+- DuckDB v1.5.0 → `1.10500.0`
+- DuckDB v1.5.1 → `1.10501.0`
+- DuckDB v1.6.0 → `1.10600.0`
+
+Verify the crate version exists on crates.io before proceeding:
+```
+https://crates.io/api/v1/crates/duckdb/versions
+```
+
+---
+
+## Step 2 — Update All Version Pins
+
+There are **7 files** with version constants to update. Update all of them:
+
+### pg_ducklake commit (`PGDUCKLAKE_COMMIT`)
+
+| File | Variable | Format |
+|------|----------|--------|
+| `Makefile` | `PGDUCKLAKE_COMMIT  ?= <hash>` | Make variable default |
+| `docker/Dockerfile` | `ARG PGDUCKLAKE_COMMIT=<hash>` | Docker ARG default (line 2) |
+| `docker/Dockerfile.daemon` | `ARG PGDUCKLAKE_COMMIT=<hash>` | Docker ARG default |
+| `docker/docker-bake.hcl` | `default = "<hash>"` | HCL variable default (in the `PGDUCKLAKE_COMMIT` variable block) |
+
+### DuckDB version (`DUCKDB_VERSION`)
+
+| File | Variable | Format |
+|------|----------|--------|
+| `docker/.env` | `DUCKDB_VERSION=v<version>` | Env var (e.g. `v1.5.1`) |
+| `docker/Dockerfile.daemon` | `ARG DUCKDB_VERSION=v<version>` | Docker ARG default |
+| `docker/docker-bake.hcl` | `default = "v<version>"` | HCL variable default (in the `DUCKDB_VERSION` variable block) |
+
+### duckdb-rs crate version
+
+| File | Variable | Format |
+|------|----------|--------|
+| `Cargo.toml` | `duckdb = { version = "=<crate_version>" }` | Exact pin in workspace deps |
+
+After editing `Cargo.toml`, update `Cargo.lock`:
+```bash
+cargo update -p duckdb
+```
+
+---
+
+## Step 3 — Ask User What Else to Do
+
+After updating all version pins, use AskUserQuestion to ask the user what additional steps to run:
+
+- Question: "Version pins updated. What else should I do?"
+- Options:
+  1. **Done** — Stop here (just the version pin updates)
+  2. **Build** — Also rebuild pg_duckpipe (`make build`)
+  3. **Build + Test** — Rebuild and run regression tests (`make installcheck`)
+
+If the user selects **Done**, skip to Step 6 (Summary).
+
+---
+
+## Step 4 — Build pg_duckpipe (if requested)
+
+```bash
+PG_CONFIG=<pg_config_path> make build
+```
+
+If there's a local pg_ducklake checkout that needs updating first, check if it's already at the target commit. If not, warn the user they may need to rebuild pg_ducklake first (via `/build` or manually).
+
+---
+
+## Step 5 — Run Regression Tests (if requested)
+
+```bash
+PG_CONFIG=<pg_config_path> make check-regression
+```
+
+All tests must pass before considering the bump complete.
+
+---
+
+## Step 6 — Summary
+
+Print a summary of what was changed:
+
+```
+Bumped pg_ducklake to <commit> (DuckDB v<version>)
+
+Files updated:
+  Makefile                  PGDUCKLAKE_COMMIT → <short_hash>
+  docker/Dockerfile         PGDUCKLAKE_COMMIT → <short_hash>
+  docker/Dockerfile.daemon  PGDUCKLAKE_COMMIT → <short_hash>, DUCKDB_VERSION → v<version>
+  docker/docker-bake.hcl    PGDUCKLAKE_COMMIT → <short_hash>, DUCKDB_VERSION → v<version>
+  docker/.env               DUCKDB_VERSION → v<version>
+  Cargo.toml                duckdb → =<crate_version>
+  Cargo.lock                (auto-updated)
+```

Cargo.lock

@@ -12,7 +12,7 @@ tracing-subscriber = { version = "0.3", features = ["env-filter"] }
 clap = { version = "4", features = ["derive", "env"] }
 tokio = { version = "1", features = ["rt", "time", "sync", "macros"] }
 tokio-postgres = { version = "0.7", features = ["with-chrono-0_4"] }
-duckdb = { version = "=1.10500.0" }
+duckdb = { version = "=1.10501.0" }
 pgwire-replication = { version = "0.2", features = ["tls-rustls"] }
 futures-util = "0.3"
 rustls = { version = "0.23", default-features = false, features = ["ring"] }

Cargo.toml

@@ -28,7 +28,7 @@ build: check-cargo-pgrx
 # Build ducklake.duckdb_extension and install to pkglibdir.
 # Skips if already present. Clones pg_ducklake on demand if DUCKLAKE_REPO is not set.
 DUCKLAKE_GIT_URL ?= https://github.com/relytcloud/pg_ducklake.git
-DUCKLAKE_COMMIT  ?= c6b2fd9408431cd86f720fba276774d2bcf23eff
+PGDUCKLAKE_COMMIT  ?= d5dd63073b9512b05f8a7ee6ebd79ff5b50b1028
 
 install-ducklake-ext:
 	@if [ -f "$(PG_LIB)/ducklake.duckdb_extension" ] && [ "$${FORCE:-0}" != "1" ]; then \
@@ -39,11 +39,11 @@ install-ducklake-ext:
 			if [ -d deps/pg_ducklake ]; then \
 				echo "==> Using existing deps/pg_ducklake"; \
 			else \
-				echo "==> Cloning pg_ducklake @ $(DUCKLAKE_COMMIT) into deps/ ..."; \
+				echo "==> Cloning pg_ducklake @ $(PGDUCKLAKE_COMMIT) into deps/ ..."; \
 				mkdir -p deps && \
 				git init deps/pg_ducklake && \
 				git -C deps/pg_ducklake remote add origin "$(DUCKLAKE_GIT_URL)" && \
-				git -C deps/pg_ducklake fetch --depth 1 origin "$(DUCKLAKE_COMMIT)" && \
+				git -C deps/pg_ducklake fetch --depth 1 origin "$(PGDUCKLAKE_COMMIT)" && \
 				git -C deps/pg_ducklake checkout FETCH_HEAD && \
 				git -C deps/pg_ducklake submodule update --init --recursive --depth 1; \
 			fi; \

Makefile

@@ -1,3 +1,3 @@
 # Shared build configuration — sourced by Dockerfiles, Makefile, and docker-bake.hcl.
 # Keep in sync with libduckdb-sys version in Cargo.lock.
-DUCKDB_VERSION=v1.5.0
+DUCKDB_VERSION=v1.5.1

docker/.env

@@ -1,12 +1,12 @@
 ARG PG_VERSION=18
-ARG DUCKLAKE_COMMIT=c6b2fd9408431cd86f720fba276774d2bcf23eff
+ARG PGDUCKLAKE_COMMIT=d5dd63073b9512b05f8a7ee6ebd79ff5b50b1028
 
 ###############################################################################
 ### DUCKLAKE-BUILDER — clone + build pg_duckdb + libduckdb + pg_ducklake
 ###############################################################################
 FROM postgres:${PG_VERSION}-bookworm AS ducklake-builder
 ARG PG_VERSION
-ARG DUCKLAKE_COMMIT
+ARG PGDUCKLAKE_COMMIT
 
 # C++ build deps for DuckDB + pg_duckdb + pg_ducklake.
 RUN apt-get update -qq && \
@@ -31,7 +31,7 @@ RUN apt-get update -qq && \
 WORKDIR /build/pg_ducklake
 RUN git init . && \
     git remote add origin https://github.com/relytcloud/pg_ducklake.git && \
-    git fetch --depth 1 origin ${DUCKLAKE_COMMIT} && \
+    git fetch --depth 1 origin ${PGDUCKLAKE_COMMIT} && \
     git checkout FETCH_HEAD && \
     git submodule update --init --recursive --depth 1
 

docker/Dockerfile

@@ -1,15 +1,15 @@
 # _BUILDER_PG is internal — not a CI matrix dimension.
 # The daemon is PG-version-agnostic; we only borrow libduckdb.so at build time.
 ARG _BUILDER_PG=18
-ARG DUCKDB_VERSION=v1.5.0
-ARG DUCKLAKE_COMMIT=c6b2fd9408431cd86f720fba276774d2bcf23eff
+ARG DUCKDB_VERSION=v1.5.1
+ARG PGDUCKLAKE_COMMIT=d5dd63073b9512b05f8a7ee6ebd79ff5b50b1028
 
 ###############################################################################
 ### DUCKLAKE-BUILDER — clone pg_ducklake + build libduckdb.so
 ###############################################################################
 FROM postgres:${_BUILDER_PG}-bookworm AS ducklake-builder
 ARG _BUILDER_PG
-ARG DUCKLAKE_COMMIT
+ARG PGDUCKLAKE_COMMIT
 
 # Minimal C++ build deps (no clang/libclang — pgrx is not used for the daemon)
 RUN apt-get update -qq && \
@@ -27,7 +27,7 @@ RUN apt-get update -qq && \
 WORKDIR /build/pg_ducklake
 RUN git init . && \
     git remote add origin https://github.com/relytcloud/pg_ducklake.git && \
-    git fetch --depth 1 origin ${DUCKLAKE_COMMIT} && \
+    git fetch --depth 1 origin ${PGDUCKLAKE_COMMIT} && \
     git checkout FETCH_HEAD && \
     git submodule update --init --recursive --depth 1
 

docker/Dockerfile.daemon

@@ -40,8 +40,20 @@ if [ ! -d "${DUCKLAKE_SRC}" ]; then
     echo "ERROR: ${DUCKLAKE_SRC} not found. Is DUCKLAKE_REPO a pg_ducklake checkout?"
     exit 1
 fi
-if [ ! -f "${DUCKLAKE_SRC}/duckdb/CMakeLists.txt" ]; then
-    echo "ERROR: ${DUCKLAKE_SRC}/duckdb/ is empty. Run 'git submodule update --init --recursive' in pg_ducklake."
+
+# DuckDB source: ducklake's own duckdb/ submodule, or the shared copy in pg_duckdb.
+# After pg_ducklake migrated ducklake from submodule to subtree, the nested
+# duckdb submodule is no longer initialised — symlink to the shared copy.
+DUCKDB_SRC="${DUCKLAKE_SRC}/duckdb"
+SHARED_DUCKDB="${REPO}/third_party/pg_duckdb/third_party/duckdb"
+if [ ! -f "${DUCKDB_SRC}/CMakeLists.txt" ] && [ -f "${SHARED_DUCKDB}/CMakeLists.txt" ]; then
+    echo "==> Symlinking shared DuckDB source into ducklake/duckdb ..."
+    ln -sfn "${SHARED_DUCKDB}" "${DUCKDB_SRC}"
+fi
+
+if [ ! -f "${DUCKDB_SRC}/CMakeLists.txt" ]; then
+    echo "ERROR: DuckDB source not found at ${DUCKDB_SRC} or ${SHARED_DUCKDB}."
+    echo "Run 'git submodule update --init --recursive' in pg_ducklake."
     exit 1
 fi
 cd "${DUCKLAKE_SRC}"