Refine and extend synthetic benchmarks stressing CDATA, text and attribute value processing.

Author: adamreichold

benches/.gitattributes

@@ -1 +1,3 @@
 cdata.xml -text
+text.xml -text
+attributes.xml -text

benches/attributes.py

@@ -0,0 +1,34 @@
+import string
+import random
+
+letters = [letter for letter in string.ascii_letters]
+spaces = ["\t", "\r\n", "\n"]
+codes = [""", "&", "'", ">", "	", "
", "
"]
+refs = [f"&entity{idx};" for idx in range(5)]
+chunks = 5 * letters + spaces + codes
+chunks_with_refs = chunks + refs
+
+with open("attributes.xml", "w") as file:
+    file.write("<!DOCTYPE test [\n")
+
+    for idx in range(5):
+        file.write(f"<!ENTITY entity{idx} '")
+
+        for _ in range(200):
+            file.write(random.choice(chunks))
+
+        file.write("'>")
+
+    file.write("]>\n\n")
+
+    file.write("<root>\n")
+
+    for _ in range(1_000):
+        file.write('\t<elem attr="')
+
+        for _ in range(200):
+            file.write(random.choice(chunks_with_refs))
+
+        file.write('"/>\n')
+
+    file.write("</root>\n")

benches/attributes.xml

@@ -0,0 +1,14 @@
+import string
+import random
+
+letters = [letter for letter in string.ascii_letters]
+lines = ["\r\n", "\n"]
+chunks = letters + lines
+
+with open("cdata.xml", "w") as file:
+    file.write("<root><![CDATA[")
+
+    for _ in range(100_000):
+        file.write(random.choice(chunks))
+
+    file.write("]]></root>")