generate XML schema docs from XML schemas themselves

1 files changed, 34 insertions, 2 deletions

diff --git a/testsuite/Testschema/test_schema.py b/testsuite/Testschema/test_schema.py
index 56fe76205..ddfe4775f 100644
--- a/testsuite/Testschema/test_schema.py
+++ b/testsuite/Testschema/test_schema.py
@@ -1,7 +1,7 @@
 import os
-import re
 import sys
 import glob
+import lxml.etree
 from subprocess import Popen, PIPE, STDOUT
 
 # add all parent testsuite directories to sys.path to allow (most)
@@ -27,14 +27,46 @@ except OSError:
     HAS_XMLLINT = False
 
 
+XS = "http://www.w3.org/2001/XMLSchema"
+XS_NS = "{%s}" % XS
+NSMAP = dict(xs=XS)
+
+
 class TestSchemas(Bcfg2TestCase):
     schema_url = "http://www.w3.org/2001/XMLSchema.xsd"
 
     @skipUnless(HAS_XMLLINT, "xmllint not installed")
     def test_valid(self):
-        schemas = [s for s in glob.glob(os.path.join(srcpath,'*.xsd'))]
+        schemas = [s for s in glob.glob(os.path.join(srcpath, '*.xsd'))]
         xmllint = Popen(['xmllint', '--xinclude', '--noout', '--schema',
                          self.schema_url] + schemas,
                         stdout=PIPE, stderr=STDOUT)
         print(xmllint.communicate()[0])
         self.assertEqual(xmllint.wait(), 0)
+
+    def test_duplicates(self):
+        entities = dict()
+        for root, _, files in os.walk(srcpath):
+            for fname in files:
+                if not fname.endswith(".xsd"):
+                    continue
+                path = os.path.join(root, fname)
+                relpath = path[len(srcpath):].strip("/")
+                schema = lxml.etree.parse(path).getroot()
+                ns = schema.get("targetNamespace")
+                if ns not in entities:
+                    entities[ns] = dict(group=dict(),
+                                        attributeGroup=dict(),
+                                        simpleType=dict(),
+                                        complexType=dict())
+                for entity in schema.xpath("//xs:*[@name]", namespaces=NSMAP):
+                    tag = entity.tag[len(XS_NS):]
+                    if tag not in entities[ns]:
+                        continue
+                    name = entity.get("name")
+                    if name in entities[ns][tag]:
+                        self.assertNotIn(name, entities[ns][tag],
+                                         "Duplicate %s %s (in %s and %s)" %
+                                         (tag, name, fname,
+                                          entities[ns][tag][name]))
+                    entities[ns][tag][name] = fname