Allow embedded zeroes (even if it pains me to do so)

project-chip · Nov 9, 2023 · 9c301c7 · 9c301c7
1 parent 967c32b
commit 9c301c7
Show file tree

Hide file tree

Showing 3 changed files with 21 additions and 16 deletions.
diff --git a/src/lib/support/tests/TestUtf8.cpp b/src/lib/support/tests/TestUtf8.cpp
@@ -66,6 +66,26 @@ void TestValidStrings(nlTestSuite * inSuite, void * inContext)
  NL_TEST_ASSERT(inSuite, IsValidCStringAsUtf8("�"));
  NL_TEST_ASSERT(inSuite, IsValidCStringAsUtf8("􏿿"));
  NL_TEST_ASSERT(inSuite, IsValidCStringAsUtf8("����"));
+
+
+ // NOTE: UTF8 allows embeded NULLs
+ // even though strings like that are probably not ideal for handling
+ // Test that we allow this, but consider later to disallow them
+ // completely if the spec is updated as such
+ {
+ char zero[16] = { 0 };
+ NL_TEST_ASSERT(inSuite, Utf8::IsValid(CharSpan(zero, 0)));
+ NL_TEST_ASSERT(inSuite, Utf8::IsValid(CharSpan(zero, 1)));
+ NL_TEST_ASSERT(inSuite, Utf8::IsValid(CharSpan(zero, 2)));
+ NL_TEST_ASSERT(inSuite, Utf8::IsValid(CharSpan(zero, 3)));
+ NL_TEST_ASSERT(inSuite, Utf8::IsValid(CharSpan(zero, 4)));
+ NL_TEST_ASSERT(inSuite, Utf8::IsValid(CharSpan(zero, 16)));
+ }
+
+ {
+ char insideZero[] = "test\0zero";
+ NL_TEST_ASSERT(inSuite, Utf8::IsValid(CharSpan(insideZero)));
+ }
 }
 
 #define TEST_INVALID_BYTES(...) \
@@ -78,9 +98,6 @@ void TestValidStrings(nlTestSuite * inSuite, void * inContext)
 
 void TestInvalidStrings(nlTestSuite * inSuite, void * inContext)
 {
- // cannot embed zeroes
- TEST_INVALID_BYTES(0x00);
-
  // overly long representation
  TEST_INVALID_BYTES(0xe0, 0b1001'1111, 0x80); // A
  TEST_INVALID_BYTES(0xed, 0b1011'0000, 0x80); // B

diff --git a/src/lib/support/utf8.cpp b/src/lib/support/utf8.cpp
@@ -55,16 +55,6 @@ enum class ParserState
 
 ParserState NextState(ParserState state, uint8_t value)
 {
- if (value == 0)
- {
- // Refuse to have embedded 0s
- //
- // NOTE: this is NOT a UTF-8 requirement but a chip encoding
- // requirement for TLV. We have this here to adhere to the
- // exposed (non-anon-namespace) API.
- return ParserState::kInvalid;
- }
-
  switch (state)
  {
  case ParserState::kFirstByte:

diff --git a/src/lib/support/utf8.h b/src/lib/support/utf8.h
@@ -23,9 +23,7 @@ namespace chip {
 namespace Utf8 {
 
 /**
- * Validate that the given span looks like a valid UTF-8 string:
- * - no embedded nulls (this is a Matter Spec requirement for encoding)
- * - valid UTF8 encoding overall
+ * Validate that the given span looks like a valid UTF-8 string
  *
  * UTF-8 encoding described at
  * https://www.unicode.org/versions/Unicode12.0.0/UnicodeStandard-12.0.pdf