Clean up repo before publishing artifacts

Make sure to clean up the repo, before publishing artifacts. Starting
from scratch is necessary if we want to avoid some edge cases that can
lead to a corrupted APT repo.

Refs freedomofpress#11

tools/publish

@@ -9,6 +9,7 @@ found in each.
 import argparse
 import logging
 import os
+import shutil
 import subprocess
 import sys
 
@@ -19,6 +20,11 @@ REPO_ROOT = os.path.abspath(os.path.join(__file__, os.pardir, os.pardir))
 REPROPRO_DIR_DEFAULT = os.path.join(REPO_ROOT, "repo")
 REPREPRO_DIR = os.environ.get("REPREPRO_BASE_DIR", REPROPRO_DIR_DEFAULT)
 REPREPRO_PUBLIC_DIR = os.path.join(REPREPRO_DIR, "public")
+REPREPRO_CLEANUP_DIRS = [
+    os.path.join(REPREPRO_DIR, "db"),
+    os.path.join(REPREPRO_PUBLIC_DIR, "pool"),
+    os.path.join(REPREPRO_PUBLIC_DIR, "dists"),
+]
 
 REPO_DIRS = [
     "dangerzone",
@@ -92,6 +98,15 @@ def validate_repo():
     logger.info("Successfully validated repo")
 
 
+def cleanup_repo():
+    """
+    Clean up previously created repo data
+    """
+    logger.info("Cleaning up repo")
+    for d in REPREPRO_CLEANUP_DIRS:
+        shutil.rmtree(d, ignore_errors=True)
+
+
 def get_release_files():
     """
     Find all Release files, to facilitate generating a detached
@@ -132,6 +147,13 @@ def parse_args():
         "--distros", required=False, nargs="+",
         help="Operate on a set of specific distros (default: all)"
     )
+    parser.add_argument(
+        "--dirty", default=False, action="store_true",
+        help=(
+            "DANGEROUS! Do not clean the APT db before publishing packages."
+            " See issue #11 on why this is important"
+        )
+    )
     parser.add_argument(
         "--TEST-no-validate", default=False, action="store_true",
         help=(
@@ -177,6 +199,16 @@ def main():
     else:
         validate_repo()
 
+    if args.distros and not args.dirty:
+        logger.error("You must specify --dirty, if you want to operate only"
+                     " on a specific set of distros")
+        return 1
+
+    if args.dirty:
+        logger.warning("Skipping repo cleanup")
+    else:
+        cleanup_repo()
+
     for repo in REPO_DIRS:
         publish_packages(repo, distros=args.distros)