Remove duplicate entries with repo priority

If a module is available from many repositories (i.e. "default" and
"subscription" are both enabled), remove duplicated entries from the
list_available() return value.

Apply the repository priority, based on the alphabetical ordering of the
repository name: "subscription" (S) has higher priority than "default"
(D).

core/imageroot/usr/local/agent/pypkg/cluster/modules.py

@@ -170,18 +170,25 @@ def get_latest_module(module, rdb):
 
 
 def list_available(rdb, skip_core_modules = False):
- modules = []
+ """Iterate over enabled repositories and return available modules respecting the repository priority."""
+ modules = {}
+ repositories = []
  # List all modules from enabled repositories
- for m in rdb.scan_iter('cluster/repository/*'):
- repo = rdb.hgetall(m)
- # Skip disabled repositories
- if int(repo["status"]) == 0:
+ for krepo in rdb.scan_iter('cluster/repository/*'):
+ repositories.append(krepo.removeprefix("cluster/repository/"))
+ # Alphabetical order, where last item has higher priority:
+ repositories.sort(reverse=True)
+ for nrepo in repositories:
+ repo = rdb.hgetall('cluster/repository/' + nrepo)
+ # Skip non-enabled repositories
+ if repo.get("status", "0") != "1":
  continue
-
- skip_testing_versions = repo["testing"] != "1"
- modules.extend(_list_repository_modules(rdb, os.path.basename(m), repo["url"], skip_core_modules, skip_testing_versions))
-
- return modules
+ skip_testing_versions = repo.get("testing", "0") != "1"
+ for rmod in _list_repository_modules(rdb, nrepo, repo["url"], skip_core_modules, skip_testing_versions):
+ if rmod["source"] in modules:
+ continue # skip duplicated images from lower priority modules
+ modules[rmod["source"]] = rmod
+ return list(modules.values())
 
 def list_installed(rdb, skip_core_modules = False):
  installed = {}