Merge pull request #20873 from github/shared-xml-discard

Share XML discard predicates

python/ql/lib/semmle/python/Overlay.qll

@@ -2,6 +2,8 @@
  * Defines entity discard predicates for Python overlay analysis.
  */
 
+private import internal.OverlayXml
+
 /*- Predicates -*/
 /**
  * Holds always for the overlay variant and never for the base variant.
@@ -303,33 +305,6 @@ final private class DiscardableComment extends Discardable instanceof @py_commen
   }
 }
 
-/*- XML -*/
-overlay[local]
-final private class DiscardableXmlLocatable extends Discardable instanceof @xmllocatable {
-  override string getPath() {
-    exists(@location loc | xmllocations(this, loc) | result = getPathForLocation(loc))
-  }
-}
-
-overlay[local]
-private predicate overlayXmlExtracted(string path) {
-  exists(DiscardableXmlLocatable d | not files(d, _) and not xmlNs(d, _, _, _) |
-    d.existsInOverlay() and
-    path = d.getPath()
-  )
-}
-
-overlay[discard_entity]
-private predicate discardXmlLocatable(@xmllocatable el) {
-  exists(DiscardableXmlLocatable d | d = el |
-    // The XML extractor is currently not incremental and may extract more
-    // XML files than those included in `overlayChangedFiles`, so this discard predicate
-    // handles those files alongside the normal `discardStarEntity` logic.
-    overlayXmlExtracted(d.getPath()) and
-    d.existsInBase()
-  )
-}
-
 /*- YAML -*/
 overlay[local]
 final private class DiscardableYamlLocatable extends Discardable instanceof @yaml_locatable {

python/ql/lib/semmle/python/internal/OverlayXml.qll

@@ -0,0 +1,46 @@
+overlay[local]
+module;
+
+/**
+ * A local predicate that always holds for the overlay variant and never holds for the base variant.
+ * This is used to define local predicates that behave differently for the base and overlay variant.
+ */
+private predicate isOverlay() { databaseMetadata("isOverlay", "true") }
+
+private string getXmlFile(@xmllocatable locatable) {
+  exists(@location_default location, @file file | xmllocations(locatable, location) |
+    locations_default(location, file, _, _, _, _) and
+    files(file, result)
+  )
+}
+
+private string getXmlFileInBase(@xmllocatable locatable) {
+  not isOverlay() and
+  result = getXmlFile(locatable)
+}
+
+/**
+ * Holds if the given `file` was extracted as part of the overlay and was extracted by the HTML/XML
+ * extractor.
+ */
+private predicate overlayXmlExtracted(string file) {
+  isOverlay() and
+  exists(@xmllocatable locatable |
+    not files(locatable, _) and not xmlNs(locatable, _, _, _) and file = getXmlFile(locatable)
+  )
+}
+
+/**
+ * Holds if the given XML `locatable` should be discarded, because it is part of the overlay base
+ * and is in a file that was also extracted as part of the overlay database.
+ */
+overlay[discard_entity]
+private predicate discardXmlLocatable(@xmllocatable locatable) {
+  exists(string file | file = getXmlFileInBase(locatable) |
+    overlayChangedFiles(file)
+    or
+    // The HTML/XML extractor is currently not incremental and may extract more files than those
+    // included in overlayChangedFiles.
+    overlayXmlExtracted(file)
+  )
+}