feat: Add configurable line_ending setting for file writes

Resolves #1139

src/serena/code_editor.py

@@ -21,11 +21,12 @@ TSymbol = TypeVar("TSymbol", bound=Symbol)
 
 
 class CodeEditor(Generic[TSymbol], ABC):
-    def __init__(self, project_root: str, project_config: ProjectConfig | None = None) -> None:
+    def __init__(self, project_root: str, project_config: ProjectConfig | None = None, newline: str | None = None) -> None:
         self.project_root = project_root
 
         # set encoding based on project configuration, if available
         self.encoding = project_config.encoding if project_config is not None else DEFAULT_SOURCE_FILE_ENCODING
+        self.newline = newline
 
     class EditedFile(ABC):
         def __init__(self, relative_path: str) -> None:
@@ -73,7 +74,7 @@ class CodeEditor(Generic[TSymbol], ABC):
     def _save_edited_file(self, edited_file: "CodeEditor.EditedFile") -> None:
         abs_path = os.path.join(self.project_root, edited_file.relative_path)
         new_contents = edited_file.get_contents()
-        with open(abs_path, "w", encoding=self.encoding) as f:
+        with open(abs_path, "w", encoding=self.encoding, newline=self.newline) as f:
             f.write(new_contents)
 
     @abstractmethod
@@ -235,8 +236,10 @@ class CodeEditor(Generic[TSymbol], ABC):
 
 
 class LanguageServerCodeEditor(CodeEditor[LanguageServerSymbol]):
-    def __init__(self, symbol_retriever: LanguageServerSymbolRetriever, project_config: ProjectConfig | None = None):
-        super().__init__(project_root=symbol_retriever.get_root_path(), project_config=project_config)
+    def __init__(
+        self, symbol_retriever: LanguageServerSymbolRetriever, project_config: ProjectConfig | None = None, newline: str | None = None
+    ):
+        super().__init__(project_root=symbol_retriever.get_root_path(), project_config=project_config, newline=newline)
         self._symbol_retriever = symbol_retriever
 
     def _get_language_server(self, relative_path: str) -> SolidLanguageServer:
@@ -372,7 +375,7 @@ class LanguageServerCodeEditor(CodeEditor[LanguageServerSymbol]):
 class JetBrainsCodeEditor(CodeEditor[JetBrainsSymbol]):
     def __init__(self, project: Project) -> None:
         self._project = project
-        super().__init__(project_root=project.project_root, project_config=project.project_config)
+        super().__init__(project_root=project.project_root, project_config=project.project_config, newline=project.line_ending.newline_str)
 
     class EditedFile(CodeEditor.EditedFile):
         def __init__(self, relative_path: str, project: Project):

src/serena/config/serena_config.py

@@ -182,6 +182,35 @@ class LanguageBackend(Enum):
         return self == LanguageBackend.JETBRAINS
 
 
+class LineEnding(Enum):
+    """Line ending convention for file writes."""
+
+    LF = "lf"
+    CRLF = "crlf"
+    NATIVE = "native"
+
+    @property
+    def newline_str(self) -> str | None:
+        """The newline parameter value for :func:`open` and :meth:`Path.write_text`.
+
+        Returns ``None`` for native mode (platform default).
+        """
+        if self is LineEnding.LF:
+            return "\n"
+        elif self is LineEnding.CRLF:
+            return "\r\n"
+        return None
+
+    @classmethod
+    def from_str(cls, value: str) -> "LineEnding":
+        """Parse a string value into a :class:`LineEnding`."""
+        try:
+            return cls(value.lower())
+        except ValueError as e:
+            valid = [le.value for le in cls]
+            raise ValueError(f"Invalid line_ending: {value!r}. Valid values are: {valid}") from e
+
+
 @dataclass
 class SharedConfig(ModeSelectionDefinition, ToolInclusionDefinition, ToStringMixin):
     """Shared between SerenaConfig and ProjectConfig, the latter used to override values in the form
@@ -191,6 +220,7 @@ class SharedConfig(ModeSelectionDefinition, ToolInclusionDefinition, ToStringMix
 
     symbol_info_budget: float | None = None
     language_backend: LanguageBackend | None = None
+    line_ending: LineEnding | None = None
     read_only_memory_patterns: list[str] = field(default_factory=list)
 
 
@@ -388,6 +418,9 @@ class ProjectConfig(SharedConfig):
         language_backend_value = data.get("language_backend")
         language_backend = LanguageBackend.from_str(language_backend_value) if language_backend_value else None
 
+        line_ending_value = data.get("line_ending")
+        line_ending = LineEnding.from_str(line_ending_value) if line_ending_value else None
+
         return cls(
             project_name=data["project_name"],
             languages=languages,
@@ -400,6 +433,7 @@ class ProjectConfig(SharedConfig):
             ignore_all_files_in_gitignore=data["ignore_all_files_in_gitignore"],
             initial_prompt=data["initial_prompt"],
             encoding=data["encoding"],
+            line_ending=line_ending,
             language_backend=language_backend,
             base_modes=data["base_modes"],
             default_modes=data["default_modes"],
@@ -413,6 +447,7 @@ class ProjectConfig(SharedConfig):
         d = dataclasses.asdict(self)
         d["languages"] = [lang.value for lang in self.languages]
         d["language_backend"] = self.language_backend.value if self.language_backend is not None else None
+        d["line_ending"] = self.line_ending.value if self.line_ending is not None else None
         return d
 
     @classmethod
@@ -595,6 +630,7 @@ class SerenaConfig(SharedConfig):
     # settings with overridden defaults
     default_modes: Sequence[str] | None = ("interactive", "editing")
     symbol_info_budget: float = 10.0
+    line_ending: LineEnding = LineEnding.NATIVE
     """
     Time budget (seconds) for requests when tools request include_info (currently
     only supported for LSP-based tools).
@@ -614,7 +650,7 @@ class SerenaConfig(SharedConfig):
     # *** static members ***
 
     CONFIG_FILE = "serena_config.yml"
-    CONFIG_FIELDS_WITH_TYPE_CONVERSION = {"projects", "language_backend"}
+    CONFIG_FIELDS_WITH_TYPE_CONVERSION = {"projects", "language_backend", "line_ending"}
 
     # *** methods ***
     @classmethod
@@ -746,6 +782,13 @@ class SerenaConfig(SharedConfig):
                 del loaded_commented_yaml["jetbrains"]
         instance.language_backend = language_backend
 
+        # determine line ending
+        line_ending = get_dataclass_default(SerenaConfig, "line_ending")
+        line_ending_value = loaded_commented_yaml.get("line_ending")
+        if line_ending_value:
+            line_ending = LineEnding.from_str(line_ending_value)
+        instance.line_ending = line_ending
+
         # migrate deprecated "gui_log_level" field if necessary
         if "gui_log_level" in loaded_commented_yaml:
             num_migrations += 1
@@ -913,6 +956,9 @@ class SerenaConfig(SharedConfig):
         # convert language backend to string
         commented_yaml["language_backend"] = self.language_backend.value
 
+        # convert line ending to string
+        commented_yaml["line_ending"] = self.line_ending.value
+
         # transfer comments from the template file
         # NOTE: The template file now uses leading comments, but we previously used trailing comments,
         #       so we apply a conversion, which detects the old style and transforms it.

src/serena/project.py

@@ -249,6 +249,9 @@ class Project(ToStringMixin):
         read_only_memory_patterns = serena_config.read_only_memory_patterns + project_config.read_only_memory_patterns
         self.memories_manager = MemoriesManager(self._serena_data_folder, read_only_memory_patterns=read_only_memory_patterns)
 
+        # resolve line ending (project -> global)
+        self.line_ending = project_config.line_ending or serena_config.line_ending
+
         self.language_server_manager: LanguageServerManager | None = None
         self._is_newly_created = is_newly_created
 

src/serena/resources/project.template.yml

@@ -31,6 +31,9 @@ languages: ["python"]
 # For a list of possible encodings, see https://docs.python.org/3.11/library/codecs.html#standard-encodings
 encoding: "utf-8"
 
+# line ending convention for file writes: "lf", "crlf", or "native" (platform default)
+line_ending: "native"
+
 # The language backend to use for this project.
 # If not set, the global setting from serena_config.yml is used.
 # Valid values: LSP, JetBrains
@@ -126,4 +129,4 @@ symbol_info_budget:
 
 # list of regex patterns which, when matched, mark a memory entry as read‑only.
 # Extends the list from the global configuration, merging the two lists.
-read_only_memory_patterns: []
+read_only_memory_patterns: []

src/serena/resources/serena_config.template.yml

@@ -7,6 +7,11 @@
 #      in your IDE).
 language_backend: LSP
 
+# line ending convention for file writes across all projects.
+# Possible values: "lf" (Unix), "crlf" (Windows), "native" (platform default).
+# Can be overridden per project in project.yml.
+line_ending: native
+
 # whether to open a graphical window with Serena's logs.
 # This is mainly supported on Windows and (partly) on Linux; not available on macOS.
 # If you prefer a browser-based tool, use the `web_dashboard` option instead.

src/serena/tools/file_tools.py

@@ -73,7 +73,7 @@ class CreateTextFileTool(Tool, ToolMarkerCanEdit):
             ), f"Cannot create file outside of the project directory, got {relative_path=}"
 
         abs_path.parent.mkdir(parents=True, exist_ok=True)
-        abs_path.write_text(content, encoding=self.project.project_config.encoding)
+        abs_path.write_text(content, encoding=self.project.project_config.encoding, newline=self.project.line_ending.newline_str)
         answer = f"File created: {relative_path}."
         if will_overwrite_existing:
             answer += " Overwrote existing file."

src/serena/tools/tools_base.py

@@ -63,7 +63,11 @@ class Component(ABC):
 
         match self.agent.get_language_backend():
             case LanguageBackend.LSP:
-                return LanguageServerCodeEditor(self.create_language_server_symbol_retriever(), project_config=self.project.project_config)
+                return LanguageServerCodeEditor(
+                    self.create_language_server_symbol_retriever(),
+                    project_config=self.project.project_config,
+                    newline=self.project.line_ending.newline_str,
+                )
             case LanguageBackend.JETBRAINS:
                 return JetBrainsCodeEditor(project=self.project)
             case _: