[orca] Filter out query strings from presentation of web content

[Joanmarie Diggs <joanied src gnome org>]

commit e6ec039c953b5a414f1c38493c8b951715bef834
Author: Joanmarie Diggs <jdiggs igalia com>
Date:   Wed Jun 12 17:21:27 2019 -0400

    Filter out query strings from presentation of web content
    
    Wikipedia references include accessible objects whose entire content
    is a giant query string. While they claim no size, the giant query
    string is exposed as an author-provided accessible name causing us to
    ignore the lack of size.

 src/orca/scripts/web/braille_generator.py |  3 +++
 src/orca/scripts/web/script_utilities.py  | 20 ++++++++++++++++++--
 src/orca/scripts/web/speech_generator.py  |  3 +++
 3 files changed, 24 insertions(+), 2 deletions(-)
---
diff --git a/src/orca/scripts/web/braille_generator.py b/src/orca/scripts/web/braille_generator.py
index 51840782f..0c8c12461 100644
--- a/src/orca/scripts/web/braille_generator.py
+++ b/src/orca/scripts/web/braille_generator.py
@@ -142,6 +142,9 @@ class BrailleGenerator(braille_generator.BrailleGenerator):
         if self._script.utilities.preferDescriptionOverName(obj):
             return [obj.description]
 
+        if not self._script.utilities.hasValidName(obj):
+            return []
+
         return super()._generateName(obj, **args)
 
     def _generateExpandedEOCs(self, obj, **args):
diff --git a/src/orca/scripts/web/script_utilities.py b/src/orca/scripts/web/script_utilities.py
index f551caffc..526250d9a 100644
--- a/src/orca/scripts/web/script_utilities.py
+++ b/src/orca/scripts/web/script_utilities.py
@@ -3250,6 +3250,21 @@ class Utilities(script_utilities.Utilities):
         self._isUselessImage[hash(obj)] = rv
         return rv
 
+    def hasValidName(self, obj):
+        if not obj.name:
+            return False
+
+        if len(obj.name.split()) > 1:
+            return True
+
+        parsed = urllib.parse.parse_qs(obj.name)
+        if len(parsed) > 2:
+            msg = "WEB: name of %s is suspected query string" % obj
+            debug.println(debug.LEVEL_INFO, msg, True)
+            return False
+
+        return True
+
     def isUselessEmptyElement(self, obj):
         if not (obj and self.inDocumentContent(obj)):
             return False
@@ -3273,9 +3288,10 @@ class Utilities(script_utilities.Utilities):
             rv = False
         elif state.contains(pyatspi.STATE_EDITABLE):
             rv = False
-        elif obj.name or obj.description or obj.childCount:
+        elif self.hasValidName(obj) or obj.description or obj.childCount:
             rv = False
-        elif "Text" in interfaces and obj.queryText().characterCount:
+        elif "Text" in interfaces and obj.queryText().characterCount \
+             and obj.queryText().getText(0, -1) != obj.name:
             rv = False
         elif "Action" in interfaces and self._getActionNames(obj):
             rv = False
diff --git a/src/orca/scripts/web/speech_generator.py b/src/orca/scripts/web/speech_generator.py
index 9d1050e24..33c8e2e42 100644
--- a/src/orca/scripts/web/speech_generator.py
+++ b/src/orca/scripts/web/speech_generator.py
@@ -176,6 +176,9 @@ class SpeechGenerator(speech_generator.SpeechGenerator):
            and not self._script.utilities.isDPub(obj):
             return []
 
+        if not self._script.utilities.hasValidName(obj):
+            return []
+
         if obj.parent and obj.name and obj.name == obj.parent.name \
            and obj != orca_state.locusOfFocus:
             return []