Merge pull request 'Make parser reject state sticky after WeaselJson_REJECT' (#15) from weaselbot/weaseljson:weaselbot/issue-12 into main

Reviewed-on: weaselab/weaseljson#15

src/parser3.h

@@ -139,6 +139,7 @@ struct Parser3 {
     stackPtr = stack();
     std::ignore = push({N_VALUE, N_WHITESPACE, T_EOF});
     inKey = false;
+    rejected = false;
     utf8Codepoint = 0;
     utf16Surrogate = 0;
     minCodepoint = 0;
@@ -161,6 +162,7 @@ struct Parser3 {
   NumDfa numDfa;
   Utf8Dfa strDfa;
   bool inKey = false;
+  bool rejected = false;
 
 #ifndef HAS_MUSTTAIL
   char *stashBufForTrampoline;
@@ -1062,18 +1064,29 @@ constexpr inline struct ContinuationTable {
 inline WeaselJsonStatus Parser3::parse(char *buf, int len) {
   this->dataBegin = this->writeBuf = buf;
 
+  if (this->rejected) [[unlikely]] {
+    return WeaselJson_REJECT;
+  }
+
 #ifdef HAS_MUSTTAIL
   // The continuation returns a value in 0..3 here (kBounce is only used by the
   // no-musttail trampoline below), so the conversion back to the enum is in
   // range.
-  return WeaselJsonStatus(
+  ContinuationStatus status =
-      symbolTables.continuations[top()](this, buf, buf + len));
+      symbolTables.continuations[top()](this, buf, buf + len);
+  if (status == WeaselJson_REJECT) {
+    this->rejected = true;
+  }
+  return WeaselJsonStatus(status);
 #else
   this->stashBufForTrampoline = buf;
   ContinuationStatus result;
   while ((result = symbolTables.continuations[top()](
               this, stashBufForTrampoline, buf + len)) == kBounce)
     ;
+  if (result == WeaselJson_REJECT) {
+    this->rejected = true;
+  }
   return WeaselJsonStatus(result);
 #endif
 }

src/test.cpp

@@ -202,6 +202,31 @@ TEST_CASE("parser3") {
   }
 }
 
+TEST_CASE("rejected state is sticky") {
+  auto c = noopCallbacks();
+  auto *parser = WeaselJsonParser_create(1024, &c, nullptr, 0);
+  REQUIRE(parser != nullptr);
+
+  // Feed a valid prefix, then an invalid byte. The parser must reject.
+  std::string chunk1 = "1";
+  REQUIRE(WeaselJsonParser_parse(parser, chunk1.data(), chunk1.size()) ==
+          WeaselJson_AGAIN);
+  std::string chunk2 = "x";
+  REQUIRE(WeaselJsonParser_parse(parser, chunk2.data(), chunk2.size()) ==
+          WeaselJson_REJECT);
+
+  // After a reject, every later call must also reject, including the EOF
+  // finish call that would otherwise return OK.
+  REQUIRE(WeaselJsonParser_parse(parser, nullptr, 0) == WeaselJson_REJECT);
+
+  // Further data chunks must also stay rejected.
+  std::string chunk3 = " ";
+  REQUIRE(WeaselJsonParser_parse(parser, chunk3.data(), chunk3.size()) ==
+          WeaselJson_REJECT);
+
+  WeaselJsonParser_destroy(parser);
+}
+
 TEST_CASE("create rejects too-small stack") {
   auto c = noopCallbacks();
   // The parser needs room for the bootstrap symbols pushed by reset(); a stack