From ee081ca70322cd01d0642fdb1860e6bfbfb48e85 Mon Sep 17 00:00:00 2001
From: steven-vd <steven@vandorp.lu>
Date: Sun, 16 Aug 2026 15:00:32 +0200
Subject: Initial commit

---
 benchmarks/RapidXML/pubmed_len.cpp | 46 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 46 insertions(+)
 create mode 100644 benchmarks/RapidXML/pubmed_len.cpp

(limited to 'benchmarks/RapidXML/pubmed_len.cpp')

diff --git a/benchmarks/RapidXML/pubmed_len.cpp b/benchmarks/RapidXML/pubmed_len.cpp
new file mode 100644
index 0000000..f25f33b
--- /dev/null
+++ b/benchmarks/RapidXML/pubmed_len.cpp
@@ -0,0 +1,46 @@
+#include "rapidxml.hpp"
+#include <cstring>
+
+inline bool is_content_root(const char* name, std::size_t len) {
+    return (len == 8 && std::memcmp(name, "abstract", 8) == 0) ||
+           (len == 4 && std::memcmp(name, "body", 4) == 0);
+}
+
+size_t sum_text(const rapidxml::xml_node<>* node) {
+    size_t total = 0;
+    for (auto* child = node->first_node(); child; child = child->next_sibling()) {
+        if (child->type() == rapidxml::node_type::node_data) {
+            total += child->value_size();
+        } else if (child->type() == rapidxml::node_type::node_element) {
+            total += sum_text(child);
+        }
+    }
+    return total;
+}
+
+size_t walk(const rapidxml::xml_node<>* node) {
+    size_t total = 0;
+    for (auto* child = node->first_node(); child; child = child->next_sibling()) {
+        if (child->type() != rapidxml::node_type::node_element) continue;
+        if (is_content_root(child->name(), child->name_size())) {
+            total += sum_text(child);
+        } else {
+            total += walk(child);
+        }
+    }
+    return total;
+}
+
+size_t pubmed_len(char* buffer, size_t /*len*/) {
+    rapidxml::xml_document<> doc;
+    doc.parse<rapidxml::parse_no_entity_translation |
+              rapidxml::parse_no_string_terminators>(buffer);
+    return walk(&doc);
+}
+
+size_t pubmed_len_with_entities(char* buffer, size_t /*len*/) {
+    rapidxml::xml_document<> doc;
+    doc.parse<rapidxml::parse_no_string_terminators>(buffer);
+    return walk(&doc);
+}
+
-- 
cgit v1.2.3