re-order conditions

adbar · Nov 8, 2023 · 140c121 · 140c121
1 parent fcb6a97
commit 140c121
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 2 deletions.
diff --git a/trafilatura/feeds.py b/trafilatura/feeds.py
@@ -38,7 +38,7 @@ def handle_link_list(linklist, domainname, baseurl, target_lang=None, external=F
         # control output for validity
         checked = check_url(link, language=target_lang)
         if checked is not None:
-            if not external and not is_similar_domain(domainname, checked[1]) and not "feed" in link:
+            if not external and not "feed" in link and not is_similar_domain(domainname, checked[1]):
                 LOGGER.warning('Rejected, diverging domain names: %s %s', domainname, checked[1])
             else:
                 output_links.append(checked[0])

diff --git a/trafilatura/sitemaps.py b/trafilatura/sitemaps.py
@@ -79,7 +79,8 @@ def handle_link(self, link: str) -> None:
 
         # don't take links from another domain and make an exception for main platforms
         # also bypass: subdomains vs. domains
-        if not self.external and not is_similar_domain(self.domain, newdomain) and not WHITELISTED_PLATFORMS.search(newdomain):
+        if not self.external  and not WHITELISTED_PLATFORMS.search(newdomain) and \
+            not is_similar_domain(self.domain, newdomain):
             LOGGER.warning('link discarded, diverging domain names: %s %s', self.domain, newdomain)
             return