diff --git a/strix/interface/utils.py b/strix/interface/utils.py index e051a185..f82dcdc9 100644 --- a/strix/interface/utils.py +++ b/strix/interface/utils.py @@ -814,6 +814,7 @@ def _classify_diff_entries(entries: list[DiffEntry]) -> dict[str, Any]: analyzable_files: list[str] = [] analyzable_seen: set[str] = set() modified_seen: set[str] = set() + added_seen: set[str] = set() for entry in entries: path = entry.path @@ -825,7 +826,7 @@ def _classify_diff_entries(entries: list[DiffEntry]) -> dict[str, Any]: continue if entry.status == "A": - added_files.append(path) + _append_unique(added_files, added_seen, path) _append_unique(analyzable_files, analyzable_seen, path) continue @@ -848,7 +849,7 @@ def _classify_diff_entries(entries: list[DiffEntry]) -> dict[str, Any]: continue if entry.status == "C": - _append_unique(modified_files, modified_seen, path) + _append_unique(added_files, added_seen, path) _append_unique(analyzable_files, analyzable_seen, path) continue @@ -1283,6 +1284,7 @@ def derive_local_base_name(path_str: str) -> str: def assign_workspace_subdirs(targets_info: list[dict[str, Any]]) -> None: name_counts: dict[str, int] = {} + allocated_subdirs: set[str] = set() for target in targets_info: target_type = target["type"] @@ -1298,10 +1300,22 @@ def assign_workspace_subdirs(targets_info: list[dict[str, Any]]) -> None: continue count = name_counts.get(base_name, 0) + 1 - name_counts[base_name] = count - workspace_subdir = base_name if count == 1 else f"{base_name}-{count}" + # Avoid colliding with a subdir already allocated to another target + # (e.g. base names ["api-2", "api", "api"] would otherwise both map to "api-2"). + # Track allocated subdirs separately from base-name counts so a derived + # suffix never poisons the count of a later target sharing that name. + max_attempts = len(targets_info) + 1 + attempts = 0 + while workspace_subdir in allocated_subdirs and attempts < max_attempts: + count += 1 + workspace_subdir = f"{base_name}-{count}" + attempts += 1 + + name_counts[base_name] = count + allocated_subdirs.add(workspace_subdir) + details["workspace_subdir"] = workspace_subdir diff --git a/tests/test_diff_workspace_utils.py b/tests/test_diff_workspace_utils.py new file mode 100644 index 00000000..2074df14 --- /dev/null +++ b/tests/test_diff_workspace_utils.py @@ -0,0 +1,62 @@ +"""Tests for diff classification and workspace-subdir assignment in interface.utils.""" + +from __future__ import annotations + +from typing import Any + +from strix.interface.utils import ( + DiffEntry, + _classify_diff_entries, + assign_workspace_subdirs, +) + + +def _local_target(path: str) -> dict[str, Any]: + return {"type": "local_code", "details": {"target_path": path}} + + +def test_assign_workspace_subdirs_avoids_suffix_collision() -> None: + # Base names derive to ["api-2", "api", "api"]; the third target's naive + # suffix "api-2" would collide with the first, silently dropping a repo. + targets = [ + _local_target("/srv/api-2"), + _local_target("/srv/api"), + _local_target("/other/api"), + ] + + assign_workspace_subdirs(targets) + + subdirs = [t["details"]["workspace_subdir"] for t in targets] + assert len(set(subdirs)) == 3 + assert subdirs[0] == "api-2" + assert subdirs[1] == "api" + assert subdirs[2] not in {"api-2", "api"} + + +def test_assign_workspace_subdirs_suffix_does_not_poison_later_base() -> None: + # Base names ["api", "api", "api-2"]: the second "api" target allocates the + # subdir "api-2"; the later real "api-2" target must still get a unique, + # non-colliding subdir without the earlier allocation poisoning its counter. + targets = [ + _local_target("/srv/api"), + _local_target("/other/api"), + _local_target("/srv/api-2"), + ] + + assign_workspace_subdirs(targets) + + subdirs = [t["details"]["workspace_subdir"] for t in targets] + assert len(set(subdirs)) == 3 + assert subdirs[0] == "api" + assert subdirs[1] == "api-2" + assert subdirs[2].startswith("api-2") + + +def test_classify_diff_entries_copy_is_added_not_modified() -> None: + result = _classify_diff_entries( + [DiffEntry(status="C", path="new.py", old_path="orig.py", similarity=100)] + ) + + assert "new.py" in result["added_files"] + assert "new.py" not in result["modified_files"] + assert "new.py" in result["analyzable_files"]