Merge pull request #6022 from tk0miya/5196_linkcheck_should_check_remote_image

Close #5196: linkcheck also checks remote images exist

CHANGES

@@ -177,6 +177,7 @@ Features added
 * #6016: HTML search: A placeholder for the search summary prevents search
   result links from changing their position when the search terminates.  This
   makes navigating search results easier.
+* #5196: linkcheck also checks remote images exist
   
 Bugs fixed
 ----------

sphinx/builders/linkcheck.py

@@ -25,7 +25,7 @@ from sphinx.util import encode_uri, requests, logging
 from sphinx.util.console import (  # type: ignore
     purple, red, darkgreen, darkgray, darkred, turquoise
 )
-from sphinx.util.nodes import traverse_parent
+from sphinx.util.nodes import get_node_line
 from sphinx.util.requests import is_ssl_error
 
 if False:
@@ -271,17 +271,24 @@ class CheckExternalLinksBuilder(Builder):
         # type: (str, nodes.Node) -> None
         logger.info('')
         n = 0
-        for node in doctree.traverse(nodes.reference):
+
-            if 'refuri' not in node:
+        # reference nodes
+        for refnode in doctree.traverse(nodes.reference):
+            if 'refuri' not in refnode:
                 continue
-            uri = node['refuri']
+            uri = refnode['refuri']
-            lineno = None
+            lineno = get_node_line(refnode)
-            for parent in traverse_parent(node):
-                if parent.line:
-                    lineno = parent.line
-                    break
             self.wqueue.put((uri, docname, lineno), False)
             n += 1
+
+        # image nodes
+        for imgnode in doctree.traverse(nodes.image):
+            uri = imgnode['candidates'].get('?')
+            if uri and '://' in uri:
+                lineno = get_node_line(imgnode)
+                self.wqueue.put((uri, docname, lineno), False)
+                n += 1
+
         done = 0
         while done < n:
             self.process_result(self.rqueue.get())

sphinx/util/nodes.py

@@ -285,6 +285,14 @@ def find_source_node(node):
     return None
 
 
+def get_node_line(node):
+    # type: (nodes.Element) -> int
+    for pnode in traverse_parent(node):
+        if pnode.line:
+            return pnode.line
+    return None
+
+
 def traverse_parent(node, cls=None):
     # type: (nodes.Element, Any) -> Iterable[nodes.Element]
     while node:

tests/roots/test-linkcheck/links.txt

@@ -11,3 +11,6 @@ Some additional anchors to exercise ignore code
 * `Example Bar invalid <http://example.com/#top>`_
 * `Example anchor invalid <http://www.sphinx-doc.org/en/1.7/intro.html#does-not-exist>`_
 * `Complete nonsense <https://localhost:7777/doesnotexist>`_
+
+.. image:: http://example.com/image.png
+.. figure:: http://example.com/image2.png

tests/test_build_linkcheck.py

@@ -24,7 +24,10 @@ def test_defaults(app, status, warning):
     assert "Anchor 'does-not-exist' not found" in content
     # looking for non-existent URL should fail
     assert " Max retries exceeded with url: /doesnotexist" in content
-    assert len(content.splitlines()) == 3
+    # images should fail
+    assert "Not Found for url: http://example.com/image.png" in content
+    assert "Not Found for url: http://example.com/image2.png" in content
+    assert len(content.splitlines()) == 5
 
 
 @pytest.mark.sphinx(
@@ -32,7 +35,9 @@ def test_defaults(app, status, warning):
     confoverrides={'linkcheck_anchors_ignore': ["^!", "^top$"],
                    'linkcheck_ignore': [
                        'https://localhost:7777/doesnotexist',
-                       'http://www.sphinx-doc.org/en/1.7/intro.html#']
+                       'http://www.sphinx-doc.org/en/1.7/intro.html#',
+                       'http://example.com/image.png',
+                       'http://example.com/image2.png']
                    })
 def test_anchors_ignored(app, status, warning):
     app.builder.build_all()