Path: Add IsFileNameValid()

src/common-tests/path_tests.cpp

@@ -6,6 +6,10 @@
 
 #include <gtest/gtest.h>
 
+#include <string_view>
+
+using namespace std::string_view_literals;
+
 TEST(Path, ToNativePath)
 {
   ASSERT_EQ(Path::ToNativePath(""), "");
@@ -234,6 +238,7 @@ TEST(Path, SanitizeFileName)
   ASSERT_EQ(Path::SanitizeFileName("abcdefghijlkmnopqrstuvwxyz-0123456789+&=_[]{}"),
             "abcdefghijlkmnopqrstuvwxyz-0123456789+&=_[]{}");
   ASSERT_EQ(Path::SanitizeFileName("some*path**with*asterisks"), "some_path__with_asterisks");
+  ASSERT_EQ(Path::SanitizeFileName("foo\0bar"sv), "foo_bar");
 #ifdef _WIN32
   ASSERT_EQ(Path::SanitizeFileName("foo:"), "foo_");
   ASSERT_EQ(Path::SanitizeFileName("foo:bar."), "foo_bar_");
@@ -244,6 +249,31 @@ TEST(Path, SanitizeFileName)
   ASSERT_EQ(Path::SanitizeFileName("foo/bar", false), "foo/bar");
 }
 
+TEST(Path, IsFileNameValid)
+{
+  ASSERT_TRUE(Path::IsFileNameValid("foo"sv));
+  ASSERT_TRUE(Path::IsFileNameValid("foo_bar-0123456789+&=_[]{}"sv));
+  ASSERT_TRUE(Path::IsFileNameValid("f🙃o"sv));
+  ASSERT_TRUE(Path::IsFileNameValid("ŻąłóРстуぬねのはen🍪⟑η∏☉ⴤℹ︎∩₲ ₱⟑♰⫳🐱"sv));
+  ASSERT_TRUE(Path::IsFileNameValid("foo/bar"sv, true));
+  ASSERT_TRUE(Path::IsFileNameValid("foo\\bar"sv, true));
+  ASSERT_FALSE(Path::IsFileNameValid("foo/bar"sv));
+  ASSERT_FALSE(Path::IsFileNameValid("foo\0bar"sv));
+  ASSERT_FALSE(Path::IsFileNameValid("foo\nbar"sv));
+#ifdef _WIN32
+  ASSERT_FALSE(Path::IsFileNameValid("foo\\bar"sv));
+  ASSERT_FALSE(Path::IsFileNameValid("foo:bar"sv));
+  ASSERT_FALSE(Path::IsFileNameValid("foo*bar"sv));
+  ASSERT_FALSE(Path::IsFileNameValid("foo?bar"sv));
+  ASSERT_FALSE(Path::IsFileNameValid("foo\"bar"sv));
+  ASSERT_FALSE(Path::IsFileNameValid("foo<bar"sv));
+  ASSERT_FALSE(Path::IsFileNameValid("foo>bar"sv));
+  ASSERT_FALSE(Path::IsFileNameValid("foo|bar"sv));
+  ASSERT_FALSE(Path::IsFileNameValid("foobar.txt."sv));
+  ASSERT_FALSE(Path::IsFileNameValid("foobar."sv));
+#endif
+}
+
 TEST(Path, RemoveLengthLimits)
 {
 #ifdef _WIN32

src/common/file_system.cpp

@@ -67,6 +67,10 @@ static bool IsUNCPath(const T& path)
 
 static inline bool FileSystemCharacterIsSane(char32_t c, bool strip_slashes)
 {
+  // no null bytes
+  if (c == 0)
+    return false;
+
 #ifdef _WIN32
   // https://docs.microsoft.com/en-gb/windows/win32/fileio/naming-a-file?redirectedfrom=MSDN#naming-conventions
   if ((c == U'/' || c == U'\\') && strip_slashes)
@@ -149,6 +153,26 @@ void Path::SanitizeFileName(std::string* str, bool strip_slashes /* = true */)
 #endif
 }
 
+bool Path::IsFileNameValid(std::string_view str, bool allow_slashes)
+{
+  size_t pos = 0;
+  while (pos < str.length())
+  {
+    char32_t ch;
+    pos += StringUtil::DecodeUTF8(str, pos, &ch);
+    if (!FileSystemCharacterIsSane(ch, !allow_slashes))
+      return false;
+  }
+
+#ifdef _WIN32
+  // Windows: Can't end filename with a period.
+  if (str.length() > 0 && str.back() == '.')
+    return false;
+#endif
+
+  return true;
+}
+
 std::string Path::RemoveLengthLimits(std::string_view str)
 {
   std::string ret;

src/common/path.h

@@ -36,6 +36,9 @@ void Canonicalize(std::string* path);
 std::string SanitizeFileName(std::string_view str, bool strip_slashes = true);
 void SanitizeFileName(std::string* str, bool strip_slashes = true);
 
+/// Returns true if the given filename contains any invalid characters.
+bool IsFileNameValid(std::string_view str, bool allow_slashes = false);
+
 /// Mutates the path to remove any MAX_PATH limits (for Windows).
 std::string RemoveLengthLimits(std::string_view str);
 void RemoveLengthLimits(std::string* path);