config/load(feat[duplicates]): preserve workspace roots when reading

why: Commands using load_configs were still dropping earlier workspace-root
entries, so duplicate paths silently lost repositories outside fmt/add/discover.

what:
- load configs through DuplicateAwareConfigReader and merge duplicates by
  default while logging conflicts or skipped merges
- keep an opt-out via merge_duplicates=False for callers that want legacy
  behavior
- add regression tests ensuring both paths keep expected repos and emit logs

Author: tony

src/vcspull/config.py

@@ -14,7 +14,7 @@
 from vcspull.validator import is_valid_config
 
 from . import exc
-from ._internal.config_reader import ConfigReader
+from ._internal.config_reader import ConfigReader, DuplicateAwareConfigReader
 from .util import get_config_dir, update_dict
 
 log = logging.getLogger(__name__)
@@ -241,6 +241,8 @@ def find_config_files(
 def load_configs(
     files: list[pathlib.Path],
     cwd: pathlib.Path | Callable[[], pathlib.Path] = pathlib.Path.cwd,
+    *,
+    merge_duplicates: bool = True,
 ) -> list[ConfigDict]:
     """Return repos from a list of files.
 
@@ -268,9 +270,44 @@ def load_configs(
         if isinstance(file, str):
             file = pathlib.Path(file)
         assert isinstance(file, pathlib.Path)
-        conf = ConfigReader._from_file(file)
-        assert is_valid_config(conf)
-        newrepos = extract_repos(conf, cwd=cwd)
+
+        config_content, duplicate_roots = (
+            DuplicateAwareConfigReader.load_with_duplicates(file)
+        )
+
+        if merge_duplicates:
+            (
+                config_content,
+                merge_conflicts,
+                _merge_change_count,
+                merge_details,
+            ) = merge_duplicate_workspace_roots(config_content, duplicate_roots)
+
+            for conflict in merge_conflicts:
+                log.warning("%s: %s", file, conflict)
+
+            for root_label, occurrence_count in merge_details:
+                duplicate_count = max(occurrence_count - 1, 0)
+                if duplicate_count == 0:
+                    continue
+                plural = "entry" if duplicate_count == 1 else "entries"
+                log.info(
+                    "%s: merged %d duplicate %s for workspace root '%s'",
+                    file,
+                    duplicate_count,
+                    plural,
+                    root_label,
+                )
+        elif duplicate_roots:
+            duplicate_list = ", ".join(sorted(duplicate_roots.keys()))
+            log.warning(
+                "%s: duplicate workspace roots detected (%s); keeping last occurrences",
+                file,
+                duplicate_list,
+            )
+
+        assert is_valid_config(config_content)
+        newrepos = extract_repos(config_content, cwd=cwd)
 
         if not repos:
             repos.extend(newrepos)

tests/test_config.py

@@ -2,6 +2,7 @@
 
 from __future__ import annotations
 
+import logging
 import typing as t
 
 import pytest
@@ -143,3 +144,65 @@ def test_extract_repos_injects_workspace_root(
         assert repo["workspace_root"] == expected_root
         expected_path = config.expand_dir(pl.Path(expected_root), cwd=tmp_path) / name
         assert repo["path"] == expected_path
+
+
+def _write_duplicate_config(tmp_path: pathlib.Path) -> pathlib.Path:
+    config_path = tmp_path / "config.yaml"
+    config_path.write_text(
+        (
+            "~/workspace/:\n"
+            "  alpha:\n"
+            "    repo: git+https://example.com/alpha.git\n"
+            "~/workspace/:\n"
+            "  beta:\n"
+            "    repo: git+https://example.com/beta.git\n"
+        ),
+        encoding="utf-8",
+    )
+    return config_path
+
+
+def test_load_configs_merges_duplicate_workspace_roots(
+    tmp_path: pathlib.Path,
+    caplog: pytest.LogCaptureFixture,
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    """Duplicate workspace roots are merged to keep every repository."""
+    monkeypatch.setenv("HOME", str(tmp_path))
+    caplog.set_level(logging.INFO, logger="vcspull.config")
+
+    config_path = _write_duplicate_config(tmp_path)
+
+    repos = config.load_configs([config_path], cwd=tmp_path)
+
+    repo_names = {repo["name"] for repo in repos}
+    assert repo_names == {"alpha", "beta"}
+
+    merged_messages = [message for message in caplog.messages if "merged" in message]
+    assert merged_messages, "Expected a merge log entry for duplicate roots"
+
+
+def test_load_configs_can_skip_merging_duplicates(
+    tmp_path: pathlib.Path,
+    caplog: pytest.LogCaptureFixture,
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    """The merge step can be skipped while still warning about duplicates."""
+    monkeypatch.setenv("HOME", str(tmp_path))
+    caplog.set_level(logging.WARNING, logger="vcspull.config")
+
+    config_path = _write_duplicate_config(tmp_path)
+
+    repos = config.load_configs(
+        [config_path],
+        cwd=tmp_path,
+        merge_duplicates=False,
+    )
+
+    repo_names = {repo["name"] for repo in repos}
+    assert repo_names == {"beta"}
+
+    warning_messages = [
+        message for message in caplog.messages if "duplicate" in message
+    ]
+    assert warning_messages, "Expected a warning about duplicate workspace roots"