Adding file validator

libqfieldsync/project_checker.py

@@ -1,4 +1,3 @@
-import re
 import sys
 from enum import Enum
 from pathlib import Path
@@ -9,7 +8,7 @@
 
 from libqfieldsync.layer import LayerSource, SyncAction, UnsupportedPrimaryKeyError
 from libqfieldsync.project import ProjectConfiguration, ProjectProperties
-from libqfieldsync.utils.file_utils import isascii
+from libqfieldsync.utils.file_utils import isascii, is_valid_filepath
 
 from .offline_converter import ExportType
 
@@ -236,7 +235,6 @@ def check_basemap_configuration(self) -> Optional[FeedbackResult]:
 
     def check_files_have_unsupported_characters(self) -> Optional[FeedbackResult]:
         problematic_paths = []
-        regexp = re.compile(r'[<>:"\\|?*]')
         home_path = Path(self.project.fileName()).parent
         try:
             for path in home_path.rglob("*"):
@@ -245,8 +243,9 @@ def check_files_have_unsupported_characters(self) -> Optional[FeedbackResult]:
                 if str(relative_path).startswith(".qfieldsync"):
                     continue
 
-                if regexp.search(str(relative_path.as_posix())) is not None:
+                if not is_valid_filepath(str(relative_path.as_posix())):
                     problematic_paths.append(relative_path)
+
         except FileNotFoundError:
             # long paths on windows will raise a FileNotFoundError in rglob, so we have to handle
             # that gracefully
@@ -257,8 +256,10 @@ def check_files_have_unsupported_characters(self) -> Optional[FeedbackResult]:
                 self.tr(
                     'Forbidden characters in filesystem path(s) "{}". '
                     'Please make sure there are no files and directories with "<", ">", ":", "/", "\\", "|", "?", "*" or double quotes (") characters in their path.'
+                    "and must not be reserved names like CON, PRN, AUX, NUL, etc."
                 ).format(", ".join([f'"{path}"' for path in problematic_paths]))
             )
+
         else:
             return None
 
@@ -309,6 +310,20 @@ def check_layer_has_ascii_filename(
         else:
             return None
 
+    # def check_layer_has_invalid_filename(
+    #     self, layer_source: LayerSource
+    # ) -> Optional[FeedbackResult]:
+    #     if layer_source.is_file and not is_valid_filename(layer_source.filename):
+    #         return FeedbackResult(
+    #             self.tr(
+    #                 'The layer "{}" has an invalid filename or path and was therefore removed from the cloud project.'
+    #                 'Please make sure there are no files and directories their path.'
+    #                 "It is highly recommended to rename them to ASCII encoded paths. "
+    #             ).format(layer_source.filename),
+    #         )
+    #     else:
+    #         return None
+
     def check_layer_primary_key(
         self, layer_source: LayerSource
     ) -> Optional[FeedbackResult]:

libqfieldsync/utils/file_utils.py

@@ -190,3 +190,25 @@ def isascii(filename: str) -> bool:
             return True
         except UnicodeEncodeError:
             return False
+
+def is_valid_filename(filename: str) -> bool:
+    """
+    Check if the filename is valid.
+    """
+    pattern = re.compile(
+        r'^(?!.*[<>:"/\\|?*])'
+        r"(?!(?:COM[0-9]|CON|LPT[0-9]|NUL|PRN|AUX|com[0-9]|con|lpt[0-9]|nul|prn|aux)$)"
+        r'[^\\\/:*"?<>|]{1,254}'
+        r"(?<![\s\.])$"
+    )
+    return bool(pattern.match(filename))
+
+def is_valid_filepath(path: str) -> bool:
+    """
+    Check if the entire path is valid by validating each part of the path.
+    """
+    path_obj = Path(path)
+    for part in path_obj.parts:
+        if not is_valid_filename(part):
+            return False
+    return True