Add libarchive 3.8.1; make it a diffoscope runtime dep

minimal.toml

@@ -15,3 +15,9 @@ exec = "claude --dangerously-skip-permissions"
 interactive = true
 packages = ["base", "bash", "opencode", "curl"]
 exec = "opencode"
+
+# UNCOMMITTED test rig for the libarchive PR — discard before push.
+[tasks.diffcheck]
+description = "Prove diffoscope actually runs with the new libarchive dep"
+packages = ["base", "coreutils", "diffoscope"]
+bash = "diffoscope --version && echo 'diffcheck: ok'"

packages/diffoscope/build.ncl

@@ -3,6 +3,7 @@ let base = import "../base/build.ncl" in
 let toolchain = import "../toolchain/build.ncl" in
 let uv = import "../uv/build.ncl" in
 let python = import "../python/build.ncl" in
+let libarchive = import "../libarchive/build.ncl" in
 let version = "325" in
 {
   name = "diffoscope",
@@ -18,11 +19,16 @@ let version = "325" in
   ],
   runtime_deps = [
     python,
+    # diffoscope imports python's libarchive-c at STARTUP, which dlopens the
+    # native libarchive.so — without it every invocation dies with
+    # "undefined symbol: archive_version_number" (found by repro-lab's
+    # rc-tools, which invokes every tool instead of trusting `command -v`).
+    libarchive,
   ],
 
   cmd = "./build.sh",
   outputs = {
-    diffoscope = { glob = "usr/bin/diffoscope" } | OutputBin,
+    diffoscope = { glob = "usr/bin/diffoscope*" } | OutputBin,
     site_packages = { glob = "usr/lib/python*/site-packages/**/*" } | OutputData,
   },
   attrs = {

packages/diffoscope/build.sh

@@ -5,5 +5,18 @@ tar -xof diffoscope-325.tar.gz
 cd diffoscope-325
 
 pip3 install --root $OUTPUT_DIR .
+
+# ctypes' find_library() has neither ldconfig nor gcc inside a composed
+# root, so libarchive-c cannot locate libarchive.so on its own — but it
+# honors $LIBARCHIVE. Ship the console script behind a wrapper that points
+# it at the library (overridable, harmless when unset elsewhere).
+mv "$OUTPUT_DIR/usr/bin/diffoscope" "$OUTPUT_DIR/usr/bin/diffoscope-real"
+cat > "$OUTPUT_DIR/usr/bin/diffoscope" <<'WRAP'
+#!/bin/sh
+export LIBARCHIVE="${LIBARCHIVE:-/usr/lib/libarchive.so.13}"
+exec /usr/bin/diffoscope-real "$@"
+WRAP
+chmod 755 "$OUTPUT_DIR/usr/bin/diffoscope"
+
 # TODO does not produce /usr/bin/diffoscope
 # uv pip install --system --prefix $OUTPUT_DIR/usr -r pyproject.toml --extra cmdline

packages/libarchive/build.ncl

@@ -0,0 +1,62 @@
+let { Attrs, BuildSpec, Local, OutputData, OutputLib, Source, .. } = import "minimal.ncl" in
+let bash-bootstrap = import "../bash-bootstrap/build.ncl" in
+let make = import "../make/build.ncl" in
+let sed = import "../sed/build.ncl" in
+let grep = import "../grep/build.ncl" in
+let gawk-bootstrap = import "../gawk-bootstrap/build.ncl" in
+let tar = import "../tar/build.ncl" in
+let coreutils = import "../coreutils/build.ncl" in
+let toolchain = import "../toolchain/build.ncl" in
+let glibc = import "../glibc/build.ncl" in
+let zlib = import "../zlib/build.ncl" in
+let bzip2 = import "../bzip2/build.ncl" in
+let zstd = import "../zstd/build.ncl" in
+
+let version = "3.8.1" in
+{
+  name = "libarchive",
+  build_deps = [
+    { file = "build.sh" } | Local,
+    {
+      # Upstream: https://github.com/libarchive/libarchive/releases/download/v3.8.1/libarchive-3.8.1.tar.gz
+      # STAGING: gcloud storage cp libarchive-3.8.1.tar.gz gs://minimal-staging-archives/
+      url = "gs://minimal-staging-archives/libarchive-%{version}.tar.gz",
+      sha256 = "bde832a5e3344dc723cfe9cc37f8e54bde04565bfe6f136bc1bd31ab352e9fab"
+    } | Source,
+    bash-bootstrap,
+    make,
+    sed,
+    grep,
+    gawk-bootstrap,
+    tar,
+    coreutils,
+    toolchain,
+    zlib,
+    bzip2,
+    zstd,
+  ],
+  runtime_deps = [
+    glibc,
+    zlib,
+    bzip2,
+    zstd,
+  ],
+
+  cmd = "./build.sh",
+  outputs = {
+    # Exists for diffoscope: python's libarchive-c dlopens libarchive.so.
+    libarchive = { glob = "usr/lib/libarchive.so*" } | OutputLib,
+    headers = { glob = "usr/include/archive*.h" } | OutputData,
+    pkgconfig = { glob = "usr/lib/pkgconfig/libarchive.pc" } | OutputData,
+  },
+  attrs =
+    {
+      upstream_version = version,
+      license_spdx = "BSD-2-Clause",
+      source_provenance = {
+        category = 'GithubRepo,
+        owner = "libarchive",
+        repo = "libarchive",
+      },
+    } | Attrs,
+} | BuildSpec

packages/libarchive/build.sh

@@ -0,0 +1,30 @@
+#!/bin/sh
+set -ex
+
+tar -xof libarchive-3.8.1.tar.gz
+cd libarchive-3.8.1
+
+case $(uname -m) in
+  x86_64)  MARCH="-march=x86-64-v3" ;;
+  aarch64) MARCH="-march=armv8-a" ;;
+  *)       MARCH="" ;;
+esac
+export CFLAGS="$MARCH -O2 -pipe -gno-record-gcc-switches -ffile-prefix-map=$(pwd)=/builddir"
+export LDFLAGS="-Wl,--build-id=none"
+export CXXFLAGS="${CFLAGS}"
+
+# Library only — the bsd* tools aren't the point (diffoscope needs
+# libarchive.so via python's libarchive-c) and every compressor we skip is a
+# runtime dep we don't drag in. zlib/bzip2/zstd are the pkgs we already have.
+./configure --prefix=/usr \
+  --disable-bsdtar --disable-bsdcat --disable-bsdcpio --disable-bsdunzip \
+  --with-zlib --with-bz2lib --with-zstd \
+  --without-lz4 --without-lzo2 --without-lzma \
+  --without-xml2 --without-expat \
+  --without-openssl --without-nettle \
+  --disable-acl --disable-xattr
+
+make -j"$(nproc)"
+# NOTE: no `make check` — libarchive's suite wants ACL/xattr/locale support
+# we deliberately configured out; failures there wouldn't indict the library.
+make DESTDIR="$OUTPUT_DIR" install