GitHub Publish: preserve <code> elements instead of converting them t…

…o `s

... so that GitHub renders HTML entities like > inside them instead of leaving them escaped. background: https://chat.indieweb.org/dev/2019-12-24#t1577174464779200

the html2text people have gone back and forth on how to handle this: Alir3z4/html2text#57, Alir3z4/html2text#109, etc. original bridgy publish github escaping was added in snarfed/bridgy#810.

README.md

@@ -305,6 +305,8 @@ Non-breaking changes:
   * Bug fix: extract feed image from `hfeed` correctly.
 * REST API:
   * Add HTTP `HEAD` support.
+* GitHub:
+  * Publish: preserve `<code>` tags instead of converting them to \`s so that GitHub renders HTML entities like `&gt;` inside them instead of leaving them escaped. [Background.](https://chat.indieweb.org/dev/2019-12-24#t1577174464779200)
 * The `cache` kwarg to `Source.original_post_discovery()` now has no effect. `webutil.util.follow_redirects()` has its own built in caching now.
 
 ### 2.2 - 2019-11-02

granary/github.py

@@ -175,6 +175,24 @@
 }
 REACTIONS_REST_CHARS = {char: name for name, char, in REACTIONS_REST.items()}
 
+
+# preserve <code> elements instead of converting them to `s so that GitHub
+# renders HTML entities like &gt; inside them instead of leaving them escaped.
+# background: https://chat.indieweb.org/dev/2019-12-24#t1577174464779200
+CODE_PLACEHOLDER_START = '~~~BRIDGY-CODE-TAG-START~~~'
+CODE_PLACEHOLDER_END = '~~~BRIDGY-CODE-TAG-END~~~'
+
+def tag_callback_code_placeholders(self, tag, attrs, start):
+  if tag == 'code':
+    self.o(CODE_PLACEHOLDER_START if start else CODE_PLACEHOLDER_END)
+    return True
+
+
+def replace_code_placeholders(text):
+  return text.replace(CODE_PLACEHOLDER_START, '<code>')\
+             .replace(CODE_PLACEHOLDER_END, '</code>')
+
+
 class GitHub(source.Source):
   """GitHub source class. See file docstring and Source class for details.
 
@@ -198,6 +216,7 @@ class GitHub(source.Source):
     'ignore_links': False,
     'protect_links': False,
     'use_automatic_links': False,
+    'tag_callback': tag_callback_code_placeholders,
   }
 
   def __init__(self, access_token=None):
@@ -514,9 +533,9 @@ def _create(self, obj, preview=None, include_link=source.OMIT_LINK,
         abort=True,
         error_plain='You need an in-reply-to GitHub repo, issue, PR, or comment URL.')
 
-    content = orig_content = html.escape(
+    content = orig_content = replace_code_placeholders(html.escape(
       self._content_for_create(obj, ignore_formatting=ignore_formatting),
-      quote=False)
+      quote=False))
     url = obj.get('url')
     if include_link == source.INCLUDE_LINK and url:
       content += '\n\n(Originally published at: %s)' % url

granary/tests/test_github.py

@@ -801,6 +801,26 @@ def test_create_escape_html(self):
     })
     self.assertIsNone(result.error_plain, result)
 
+  def test_create_doesnt_escape_html_inside_code_tag(self):
+    """https://github.com/indieweb/fragmention/issues/3"""
+    content = 'abc <code>&lt;div style="height: 10000px"&gt;&lt;/div&gt;</code> xyz'
+    self.expect_graphql_get_labels([])
+    self.expect_requests_post(github.REST_API_CREATE_ISSUE % ('foo', 'bar'), json={
+        'title': 'an issue title',
+        'body': content,
+        'labels': [],
+      }, response={
+        'html_url': 'https://github.com/foo/bar/issues/123',
+      }, headers=EXPECTED_HEADERS)
+    self.mox.ReplayAll()
+
+    result = self.gh.create({
+      'title': 'an issue title',
+      'content': content,
+      'inReplyTo': [{'url': 'https://github.com/foo/bar/issues'}],
+    })
+    self.assertIsNone(result.error_plain, result)
+
   def test_preview_issue(self):
     for i in range(2):
       self.expect_graphql_get_labels(['new silo'])