fix: relax collection path restriction (#340)

Author: JalinWang

python/tests/detail/test_collection_create_and_open.py

@@ -127,12 +127,11 @@ def check_collection_full(coll: Collection):
 valid_path_list = [
     "/tmp/nonexistent/directory/test_collection",
     "test/collection/with/slashes",
+    "test/collection/with/slashes/哈哈",
 ]
 invalid_path_list = [
-    "invalid:path",
+    "invalid\0path",
     "",
-    "test_collection_with_spaces ",
-    "test@#$%collection",
 ]
 
 

src/db/collection.cc

@@ -1725,9 +1725,9 @@ Status CollectionImpl::create() {
   if (path_.empty()) {
     return Status::InvalidArgument("path validate failed: path is empty");
   }
-  if (!std::regex_match(path_, COLLECTION_PATH_REGEX)) {
+  if (!FileHelper::PathSimpleValidation(path_)) {
     return Status::InvalidArgument("path validate failed: path[", path_,
-                                   "] cannot pass the regex verification");
+                                   "] is not a valid path");
   }
   if (ailego::FileHelper::IsExist(path_.c_str())) {
     return Status::InvalidArgument("path validate failed: path[", path_,

src/db/common/constants.h

@@ -46,9 +46,6 @@ const std::regex FIELD_NAME_REGEX("^[a-zA-Z0-9_-]{1,32}$");
 
 const std::regex DOC_PK_REGEX("^[a-zA-Z0-9_!@#$%+=.-]{1,64}$");
 
-const std::regex COLLECTION_PATH_REGEX(
-    R"(^(?:[a-zA-Z]:)?[/\\]?(?:[a-zA-Z0-9_.\-]+[/\\])*[a-zA-Z0-9_.\-]+$)");
-
 constexpr uint32_t kMaxDenseDimSize = 20000;
 
 constexpr uint32_t kMaxScalarFieldSize = 1024;

src/db/common/file_helper.h

@@ -231,6 +231,22 @@ class FileHelper {
     return ailego::FileHelper::FileSize(file_path.c_str());
   }
 
+  //! Perform a lightweight sanity check on the path string.
+  //! This only catches obvious invalid input and does NOT guarantee the path
+  //! is usable.
+  static bool PathSimpleValidation(const std::string &path) {
+    if (path.empty()) return false;
+
+    if (path.find('\0') != std::string::npos) return false;
+
+#ifdef _WIN32
+    // Characters forbidden in Windows path components.
+    if (path.find_first_of("<>\"|?*") != std::string::npos) return false;
+#endif
+
+    return true;
+  }
+
   //! Copy file
   //! src_file_path and dst_file_path must be the full path
   //! dst_file_path/.. must exist

tests/db/collection_test.cc

@@ -224,6 +224,7 @@ TEST_F(CollectionTest, Feature_CreateAndOpen_PathValidate) {
                                             ".hidden",
                                             "file.txt",
                                             "abs_test/nested/path",
+                                            "abs test/nested/path",
                                             "nested/a/b/c",
                                             "_",
                                             "-",
@@ -242,30 +243,15 @@ TEST_F(CollectionTest, Feature_CreateAndOpen_PathValidate) {
   }
 
   {
-    std::vector<std::string> inalid_paths = {
-        " ",         "",
-        "file name",  // space
-        "file$name",  // $
-        "a&b",        // &
-        "a|b",        // |
-        "a<b",        // <
-        "a>b",        // >
-        "a\"b",       // "
-        "a'b",        // '
-        "a;b",        // ;
-        "a?b",        // ?
-        "a*b",        // *
-        "a[b]",       // []
-        "a{b}",       // {}
-        "a~b",        // ~
-        "a#b",        // #
-        "a\tb",       // tab
-        "a\nb",       // newline
-        "a\rb",       // carriage return
+    using std::string_literals::operator""s;
+    std::vector<std::string> invalid_paths = {
+        "",
+        "v\0v"s,  // NUL
+#if _WIN32
+        "v?v"s,
+#endif
     };
-    for (auto path : inalid_paths) {
-      ailego::FileHelper::RemoveDirectory(path.c_str());
-
+    for (auto path : invalid_paths) {
       auto result = Collection::CreateAndOpen(path, *schema, options);
       if (!result.has_value()) {
         std::cout << result.error().message() << std::endl;