reproducible debian: handle new packages with multiple versions

This commit fixes a bug introduced by 07849060

Commit 07849060 changed the UPSERT that added new packages to the sources
table into an INSERT. If the new package has two versions for (suite, arch)
in debian archive, then the INSERT to sources will fail on unique constraint.
The changes in this commit provide the same functionality without the UPSERT:
there will be no failure, instead only the second source encountered in
archive will be saved.

Signed-off-by: Holger Levsen <holger@layer-acht.org>

1 files changed, 13 insertions, 2 deletions

diff --git a/bin/reproducible_scheduler.py b/bin/reproducible_scheduler.py
index c69e2909..b279e021 100755
--- a/bin/reproducible_scheduler.py
+++ b/bin/reproducible_scheduler.py
@@ -221,10 +221,21 @@ def update_sources(suite):
 
 def update_sources_db(suite, arch, sources):
     # extract relevant info (package name and version) from the sources file
-    new_pkgs = []
+    new_pkgs = set()
+    newest_version = {}
     for src in deb822.Sources.iter_paragraphs(sources.split('\n')):
         pkg = (src['Package'], src['Version'], suite, arch)
-        new_pkgs.append(pkg)
+
+        # only keep the most recent version of a src for each package/suite/arch
+        key = src['Package'] + suite + arch
+        if key in newest_version:
+            oldversion = newest_version[key]
+            oldpackage = (src['Package'], oldversion, suite, arch)
+            new_pkgs.remove(oldpackage)
+
+        newest_version[key] = src['Version']
+        new_pkgs.add(pkg)
+
     # get the current packages in the database
     query = "SELECT name, version, suite, architecture FROM sources " + \
             "WHERE suite='{}' AND architecture='{}'".format(suite, arch)