From 06c7822f791b9abf22bba607c9094ab9b713b394 Mon Sep 17 00:00:00 2001
From: dualfroz <me@dualfroz.com>
Date: Sat, 10 Oct 2026 15:34:44 +0200
Subject: [PATCH] Skip listing containers whose entries hold user cards or
 quotes

A thread without DiscussionForumPosting markup, where every poster writes
once, still passed the teaser test: each post opens with a linked name that
no other post repeats. On katzen-forum.net with the JSON-LD removed and one
post per poster, recall took the whole container (3,573 chars, seven user
cards instead of one).

Teasers hold neither a description list nor a quote, while forum posts keep
the poster's user card in one and quote each other, so an entry holding
either now rules the container out.
---
 tests/unit_tests.py           | 29 +++++++++++++++++++----------
 trafilatura/main_extractor.py |  5 ++++-
 2 files changed, 23 insertions(+), 11 deletions(-)

diff --git a/tests/unit_tests.py b/tests/unit_tests.py
index 8ebabe1..4aae421 100644
--- a/tests/unit_tests.py
+++ b/tests/unit_tests.py
@@ -2963,13 +2963,18 @@ def test_listing_page_with_links_keeps_every_entry_heading():
         assert text in result
 
 
-def _forum_thread(posters, linked=True, forum_markup=False):
+def _forum_thread(posters, linked=True, forum_markup=False, card_in_dl=False, quoting=False):
     "Thread markup: sibling <article> posts, each opening with a user card that names its poster in a heading."
     name_markup = '<a href="/members/{0}/">{0}</a>' if linked else "<span>{0}</span>"
+    card = (
+        "<dl><dt>Joined</dt><dd>March 2015</dd><dt>Messages</dt><dd>1,447</dd></dl>"
+        if card_in_dl
+        else '<div class="user-title">Joined March 2015, 1,447 messages</div>'
+    )
     posts = "".join(
-        f'<article><div class="poster"><h4>{name_markup.format(name)}</h4>'
-        "<dl><dt>Joined</dt><dd>March 2015</dd><dt>Messages</dt><dd>1,447</dd></dl></div>"
-        f"<div><p>Reply {i} from {name} carries enough prose to count as a post, running over a couple "
+        f'<article><div class="poster"><h4>{name_markup.format(name)}</h4>{card}</div>'
+        + (f"<blockquote><p>{posters[i - 1]} wrote: see my reply above.</p></blockquote>" if quoting and i else "")
+        + f"<div><p>Reply {i} from {name} carries enough prose to count as a post, running over a couple "
         "of clauses so the thread reads as a real discussion between people.</p></div></article>"
         for i, name in enumerate(posters)
     )
@@ -2986,20 +2991,24 @@ def _forum_thread(posters, linked=True, forum_markup=False):
 
 
 @pytest.mark.parametrize(
-    "posters,linked,forum_markup",
+    "posters,markup",
     [
         # the same poster comes back along the thread
-        (["alice", "bob", "alice", "carol", "bob", "alice"], True, False),
+        (["alice", "bob", "alice", "carol", "bob", "alice"], {}),
         # poster names without a link
-        (["alice", "bob", "carol", "dave", "erin", "frank"], False, False),
+        (["alice", "bob", "carol", "dave", "erin", "frank"], {"linked": False}),
         # every poster different, but the page declares itself a forum thread
-        (["alice", "bob", "carol", "dave", "erin", "frank"], True, True),
+        (["alice", "bob", "carol", "dave", "erin", "frank"], {"forum_markup": True}),
+        # every poster different and no forum markup, but the user card is a description list
+        (["alice", "bob", "carol", "dave", "erin", "frank"], {"card_in_dl": True}),
+        # every poster different and no forum markup, but the replies quote the post before them
+        (["alice", "bob", "carol", "dave", "erin", "frank"], {"quoting": True}),
     ],
 )
 @pytest.mark.parametrize("settings", [{"favor_recall": True}, {"favor_recall": True, "fast": True}])
-def test_forum_thread_is_not_a_listing(posters, linked, forum_markup, settings):
+def test_forum_thread_is_not_a_listing(posters, markup, settings):
     "forum posts line up like listing entries, but taking their container brings every user card along"
-    doc = _forum_thread(posters, linked, forum_markup)
+    doc = _forum_thread(posters, **markup)
     with _without_listing_expression():
         expected = extract(doc, output_format="markdown", config=use_config(), **settings)
     result = extract(doc, output_format="markdown", config=use_config(), **settings) or ""
diff --git a/trafilatura/main_extractor.py b/trafilatura/main_extractor.py
index 9f230ac..3cc4498 100644
--- a/trafilatura/main_extractor.py
+++ b/trafilatura/main_extractor.py
@@ -665,9 +665,12 @@ def _holds_most_of_the_text(tree: HtmlElement, expr: XPath) -> bool:
 def _entries_link_out(container: _Element) -> bool:
     """Whether every entry of the container is a teaser: its first heading links to a page of its
     own, which no other entry links to. Forum posts share the shape but open with the poster's
-    name, which comes back along the thread or carries no link at all."""
+    name, which comes back along the thread or carries no link at all, and they hold user cards
+    (description lists) and quotes, which teasers do not."""
     targets: set[str] = set()
     for entry in container.iterchildren("article"):
+        if entry.find(".//list[@rend='dl']") is not None or entry.find(".//quote") is not None:
+            return False
         heading = entry.find(".//head")
         link = heading.find(".//ref") if heading is not None else None
         if link is None: