1 files changed, 45 insertions, 13 deletions

diff --git a/absl/strings/internal/cord_internal.h b/absl/strings/internal/cord_internal.h
index 8db6aa6d..ece3db15 100644
--- a/absl/strings/internal/cord_internal.h
+++ b/absl/strings/internal/cord_internal.h
@@ -21,6 +21,7 @@
 #include <cstdint>
 #include <type_traits>
 
+#include "absl/base/attributes.h"
 #include "absl/base/config.h"
 #include "absl/base/internal/endian.h"
 #include "absl/base/internal/invoke.h"
@@ -33,6 +34,19 @@ namespace absl {
 ABSL_NAMESPACE_BEGIN
 namespace cord_internal {
 
+// The overhead of a vtable is too much for Cord, so we roll our own subclasses
+// using only a single byte to differentiate classes from each other - the "tag"
+// byte.  Define the subclasses first so we can provide downcasting helper
+// functions in the base class.
+struct CordRep;
+struct CordRepConcat;
+struct CordRepExternal;
+struct CordRepFlat;
+struct CordRepSubstring;
+struct CordRepCrc;
+class CordRepRing;
+class CordRepBtree;
+
 class CordzInfo;
 
 // Default feature enable states for cord ring buffers
@@ -74,6 +88,9 @@ enum Constants {
   kMaxBytesToCopy = 511
 };
 
+// Emits a fatal error "Unexpected node type: xyz" and aborts the program.
+ABSL_ATTRIBUTE_NORETURN void LogFatalNodeType(CordRep* rep);
+
 // Compact class for tracking the reference count and state flags for CordRep
 // instances.  Data is stored in an atomic int32_t for compactness and speed.
 class RefcountAndFlags {
@@ -156,19 +173,6 @@ class RefcountAndFlags {
   std::atomic<int32_t> count_;
 };
 
-// The overhead of a vtable is too much for Cord, so we roll our own subclasses
-// using only a single byte to differentiate classes from each other - the "tag"
-// byte.  Define the subclasses first so we can provide downcasting helper
-// functions in the base class.
-
-struct CordRepConcat;
-struct CordRepExternal;
-struct CordRepFlat;
-struct CordRepSubstring;
-struct CordRepCrc;
-class CordRepRing;
-class CordRepBtree;
-
 // Various representations that we allow
 enum CordRepKind {
   UNUSED_0 = 0,
@@ -276,6 +280,12 @@ struct CordRep {
 struct CordRepSubstring : public CordRep {
   size_t start;  // Starting offset of substring in child
   CordRep* child;
+
+  // Creates a substring on `child`, adopting a reference on `child`.
+  // Requires `child` to be either a flat or external node, and `pos` and `n` to
+  // form a non-empty partial sub range of `'child`, i.e.:
+  // `n > 0 && n < length && n + pos <= length`
+  static inline CordRepSubstring* Create(CordRep* child, size_t pos, size_t n);
 };
 
 // Type for function pointer that will invoke the releaser function and also
@@ -339,6 +349,28 @@ struct CordRepExternalImpl
   }
 };
 
+inline CordRepSubstring* CordRepSubstring::Create(CordRep* child, size_t pos,
+                                                  size_t n) {
+  assert(child != nullptr);
+  assert(n > 0);
+  assert(n < child->length);
+  assert(pos < child->length);
+  assert(n <= child->length - pos);
+
+  // TODO(b/217376272): Harden internal logic.
+  // Move to strategical places inside the Cord logic and make this an assert.
+  if (ABSL_PREDICT_FALSE(!(child->IsExternal() || child->IsFlat()))) {
+    LogFatalNodeType(child);
+  }
+
+  CordRepSubstring* rep = new CordRepSubstring();
+  rep->length = n;
+  rep->tag = SUBSTRING;
+  rep->start = pos;
+  rep->child = child;
+  return rep;
+}
+
 inline void CordRepExternal::Delete(CordRep* rep) {
   assert(rep != nullptr && rep->IsExternal());
   auto* rep_external = static_cast<CordRepExternal*>(rep);