Prune unneeded packages from the ISO mirror cache

builder/build-iso.sh

@@ -201,48 +201,49 @@ if ! download_offline_packages; then
   download_offline_packages
 fi
 
-prune_stale_package_versions() {
-  local dir="$1"
-  local pkgfile name version cmp
-  local -a stale=()
-  declare -A newest_file=()
-  declare -A newest_version=()
+# Resolve the exact filenames chosen by the same synced package databases used
+# for the download. Pruning by this transaction (rather than merely keeping the
+# newest version of every cached package name) removes packages that have left
+# the lists or dependency closure, such as an old Electron major version.
+if ! resolved_package_files="$(
+  pacman --config "/configs/pacman-online-${OMARCHY_MIRROR}.conf" --noconfirm \
+    --dbpath /tmp/offlinedb -S --print --print-format '%f' "${all_packages[@]}"
+)"; then
+  echo "ERROR: could not resolve the package files required by the offline mirror" >&2
+  exit 1
+fi
+mapfile -t required_package_files <<< "$resolved_package_files"
 
-  while IFS= read -r pkgfile; do
-    read -r name version < <(pacman -Qp "$pkgfile" 2>/dev/null) || continue
-    [[ -n $name && -n $version ]] || continue
-
-    if [[ -z ${newest_version[$name]+x} ]]; then
-      newest_version[$name]="$version"
-      newest_file[$name]="$pkgfile"
-      continue
-    fi
-
-    cmp=$(vercmp "$version" "${newest_version[$name]}")
-    if (( cmp > 0 )); then
-      stale+=("${newest_file[$name]}")
-      newest_version[$name]="$version"
-      newest_file[$name]="$pkgfile"
-    else
-      stale+=("$pkgfile")
-    fi
-  done < <(find "$dir" -maxdepth 1 -type f -name '*.pkg.tar.*' ! -name '*.sig' -print | sort)
-
-  if (( ${#stale[@]} > 0 )); then
-    echo "Pruning stale package versions from offline mirror:"
-    for pkgfile in "${stale[@]}"; do
-      echo "  $(basename "$pkgfile")"
-      rm -f "$pkgfile" "$pkgfile.sig"
+# The online transaction intentionally excludes packages built from the local
+# checkouts. Add those exact artifacts back to the keep-set after verifying
+# that the local build left exactly one file for each selected package name.
+if [[ -n ${LOCAL_OMARCHY_BUILD:-} ]]; then
+  for local_package_name in \
+    "$OMARCHY_RUNTIME_PACKAGE" "$OMARCHY_SETTINGS_PACKAGE" "$OMARCHY_NVIM_PACKAGE"; do
+    local_package_file=""
+    for candidate in "$offline_mirror_dir/$local_package_name-"*.pkg.tar.*; do
+      [[ -f $candidate && $candidate != *.sig ]] || continue
+      read -r candidate_name _ < <(pacman -Qp "$candidate" 2>/dev/null) || continue
+      [[ $candidate_name == "$local_package_name" ]] || continue
+      if [[ -n $local_package_file ]]; then
+        echo "ERROR: multiple local builds found for $local_package_name" >&2
+        exit 1
+      fi
+      local_package_file="${candidate##*/}"
     done
-  fi
-}
+    if [[ -z $local_package_file ]]; then
+      echo "ERROR: local build not found for $local_package_name" >&2
+      exit 1
+    fi
+    required_package_files+=("$local_package_file")
+  done
+fi
+
+printf '%s\n' "${required_package_files[@]}" |
+  bash /builder/prune-offline-mirror.sh "$offline_mirror_dir"
 
 # Rebuild the offline repo db from scratch so size/checksum/depends entries
-# always reflect the current package files. The persistent cache can contain
-# multiple versions of the same package; repo-add keeps the last one it sees,
-# which is lexical glob order, not version order (e.g. pkgrel -9 can override
-# -15). Prune to one newest version per package before indexing.
-prune_stale_package_versions "$offline_mirror_dir"
+# always reflect only the package files selected for this build.
 rm -f "$offline_mirror_dir"/offline.db* "$offline_mirror_dir"/offline.files*
 repo-add "$offline_mirror_dir/offline.db.tar.gz" "$offline_mirror_dir/"*.pkg.tar.zst
 

builder/prune-offline-mirror.sh

@@ -0,0 +1,67 @@
+#!/bin/bash
+
+set -euo pipefail
+
+offline_mirror_dir="${1:-}"
+if [[ -z $offline_mirror_dir ]]; then
+  echo "Usage: prune-offline-mirror.sh <offline-mirror-dir>" >&2
+  exit 1
+fi
+if [[ ! -d $offline_mirror_dir ]]; then
+  echo "ERROR: offline mirror directory not found: $offline_mirror_dir" >&2
+  exit 1
+fi
+
+# Read the exact package filenames selected by pacman on stdin. Validate the
+# complete selection before deleting anything: an empty/partial resolution must
+# never turn a warm-cache build into a broken mirror.
+declare -A required=()
+while IFS= read -r filename; do
+  [[ -n $filename ]] || continue
+  if [[ $filename == */* || $filename != *.pkg.tar.* || $filename == *.sig ]]; then
+    echo "ERROR: invalid required package filename: $filename" >&2
+    exit 1
+  fi
+  required["$filename"]=1
+done
+
+if (( ${#required[@]} == 0 )); then
+  echo "ERROR: refusing to prune the offline mirror with an empty package selection" >&2
+  exit 1
+fi
+
+missing=()
+for filename in "${!required[@]}"; do
+  [[ -f "$offline_mirror_dir/$filename" ]] || missing+=("$filename")
+done
+if (( ${#missing[@]} > 0 )); then
+  echo "ERROR: refusing to prune; selected packages are missing from the mirror:" >&2
+  printf '  %s\n' "${missing[@]}" >&2
+  exit 1
+fi
+
+stale=()
+while IFS= read -r -d '' package_file; do
+  filename="${package_file##*/}"
+  [[ -n ${required[$filename]+x} ]] || stale+=("$package_file")
+done < <(
+  find "$offline_mirror_dir" -maxdepth 1 -type f \
+    -name '*.pkg.tar.*' ! -name '*.sig' -print0
+)
+
+if (( ${#stale[@]} > 0 )); then
+  echo "Pruning package archives not selected for the offline mirror:"
+  for package_file in "${stale[@]}"; do
+    echo "  $(basename "$package_file")"
+    rm -f -- "$package_file" "$package_file.sig"
+  done
+fi
+
+# A failed or interrupted download can leave a signature without its archive.
+# It cannot be indexed and only wastes space in the ISO.
+while IFS= read -r -d '' signature_file; do
+  [[ -f ${signature_file%.sig} ]] || rm -f -- "$signature_file"
+done < <(
+  find "$offline_mirror_dir" -maxdepth 1 -type f \
+    -name '*.pkg.tar.*.sig' -print0
+)

test/test_offline_mirror_pruning.py

@@ -0,0 +1,79 @@
+#!/usr/bin/python
+
+import subprocess
+import tempfile
+import unittest
+from pathlib import Path
+
+
+ROOT = Path(__file__).resolve().parents[1]
+PRUNER = ROOT / "builder/prune-offline-mirror.sh"
+
+
+class OfflineMirrorPruningTest(unittest.TestCase):
+    def run_pruner(self, mirror: Path, required: list[str]):
+        return subprocess.run(
+            ["bash", str(PRUNER), str(mirror)],
+            input="".join(f"{filename}\n" for filename in required),
+            capture_output=True,
+            text=True,
+        )
+
+    def add_file(self, mirror: Path, filename: str):
+        (mirror / filename).write_bytes(b"package")
+
+    def test_keeps_only_exactly_selected_package_archives(self):
+        with tempfile.TemporaryDirectory() as directory:
+            mirror = Path(directory)
+            keep = "keep-2.0-1-x86_64.pkg.tar.zst"
+            other = "other-1.0-1-any.pkg.tar.zst"
+            stale_version = "keep-1.0-1-x86_64.pkg.tar.zst"
+            removed_package = "removed-1.0-1-x86_64.pkg.tar.zst"
+            orphan_signature = "interrupted-1.0-1-x86_64.pkg.tar.zst.sig"
+
+            for filename in (keep, other, stale_version, removed_package):
+                self.add_file(mirror, filename)
+            for filename in (keep, stale_version, removed_package):
+                self.add_file(mirror, f"{filename}.sig")
+            self.add_file(mirror, orphan_signature)
+            self.add_file(mirror, "offline.db.tar.gz")
+
+            result = self.run_pruner(mirror, [keep, other])
+
+            self.assertEqual(result.returncode, 0, result.stderr)
+            self.assertEqual(
+                {path.name for path in mirror.iterdir()},
+                {keep, f"{keep}.sig", other, "offline.db.tar.gz"},
+            )
+            self.assertIn(stale_version, result.stdout)
+            self.assertIn(removed_package, result.stdout)
+
+    def test_missing_selection_aborts_before_deleting_anything(self):
+        with tempfile.TemporaryDirectory() as directory:
+            mirror = Path(directory)
+            cached = "cached-1.0-1-x86_64.pkg.tar.zst"
+            self.add_file(mirror, cached)
+
+            result = self.run_pruner(
+                mirror, ["missing-1.0-1-x86_64.pkg.tar.zst"]
+            )
+
+            self.assertNotEqual(result.returncode, 0)
+            self.assertTrue((mirror / cached).exists())
+            self.assertIn("selected packages are missing", result.stderr)
+
+    def test_empty_selection_aborts_before_deleting_anything(self):
+        with tempfile.TemporaryDirectory() as directory:
+            mirror = Path(directory)
+            cached = "cached-1.0-1-x86_64.pkg.tar.zst"
+            self.add_file(mirror, cached)
+
+            result = self.run_pruner(mirror, [])
+
+            self.assertNotEqual(result.returncode, 0)
+            self.assertTrue((mirror / cached).exists())
+            self.assertIn("empty package selection", result.stderr)
+
+
+if __name__ == "__main__":
+    unittest.main()