Check if this issue is valid — if so, understand the root cause and fix it. At archivebox/misc/util.py, line 44:

<comment>Rebuilding netloc from hostname/port strips IPv6 brackets, so IPv6 URLs like http://[::1]:80 become http://::1:80, which is invalid and will break parsing/deduping.</comment>

<file context>
@@ -41,11 +41,12 @@
 path = lambda url: urlparse(url).path
 basename = lambda url: urlparse(url).path.rsplit('/', 1)[-1]
-domain = lambda url: urlparse(url).netloc
+domain = lambda url: (lambda p: f'{p.hostname}:{p.port}' if p.port else (p.hostname or p.netloc))(urlparse(url))  # returns host:port without HTTP basic auth credentials
 query = lambda url: urlparse(url).query
 fragment = lambda url: urlparse(url).fragment
</file context>

Check if this issue is valid — if so, understand the root cause and fix it. At archivebox/plugins/parse_txt_urls/on_Snapshot__71_parse_txt_urls.py, line 99:

<comment>Strict port equality prevents auth propagation between same-origin URLs when one uses an explicit default port (None vs 80/443).</comment>

<file context>
@@ -79,6 +79,38 @@ def find_all_urls(text: str):
+        return url  # Child URL already has credentials
+
+    # Only inject if same host (hostname and port match)
+    if root_parsed.hostname != child_parsed.hostname or root_parsed.port != child_parsed.port:
+        return url
+
</file context>

Check if this issue is valid — if so, understand the root cause and fix it. At archivebox/plugins/parse_txt_urls/on_Snapshot__71_parse_txt_urls.py, line 161:

<comment>Source URL can be re-added after auth injection because the skip check is performed before `inject_url_auth`, allowing the authenticated root URL to be reinserted into `urls_found`.</comment>

<file context>
@@ -125,6 +157,8 @@ def main(url: str, snapshot_id: str = None, crawl_id: str = None, depth: int = 0
         # Skip the source URL itself
         if cleaned_url != url:
+            # Propagate HTTP basic auth credentials to discovered URLs on the same host
+            cleaned_url = inject_url_auth(cleaned_url, url)
             urls_found.add(cleaned_url)
 
</file context>

Check if this issue is valid — if so, understand the root cause and fix it. At archivebox/plugins/parse_rss_urls/on_Snapshot__72_parse_rss_urls.py, line 61:

<comment>Reconstructed netloc omits required brackets for IPv6 hosts, producing invalid URLs when injecting credentials (e.g., user:pass@::1 instead of user:pass@[::1]).</comment>

<file context>
@@ -33,6 +33,38 @@
+    if root_parsed.password:
+        auth = f'{root_parsed.username}:{root_parsed.password}'
+    if child_parsed.port:
+        new_netloc = f'{auth}@{child_parsed.hostname}:{child_parsed.port}'
+    else:
+        new_netloc = f'{auth}@{child_parsed.hostname}'
</file context>

Check if this issue is valid — if so, understand the root cause and fix it. At archivebox/plugins/parse_rss_urls/on_Snapshot__72_parse_rss_urls.py, line 59:

<comment>Reinserting `urlparse`'s decoded username/password directly into the netloc can corrupt URLs when credentials contain reserved characters (e.g., `@`, `:`), changing the userinfo/host boundary and breaking authentication. The new `inject_url_auth` should re-encode userinfo before rebuilding the URL.</comment>

<file context>
@@ -33,6 +33,38 @@
+    # Rebuild netloc with credentials: user:pass@host or user:pass@host:port
+    auth = root_parsed.username
+    if root_parsed.password:
+        auth = f'{root_parsed.username}:{root_parsed.password}'
+    if child_parsed.port:
+        new_netloc = f'{auth}@{child_parsed.hostname}:{child_parsed.port}'
</file context>

Check if this issue is valid — if so, understand the root cause and fix it. At archivebox/plugins/parse_html_urls/on_Snapshot__70_parse_html_urls.py, line 114:

<comment>Auth injection can be skipped for the same endpoint when one URL includes an explicit default port and the other omits it, because urlparse().port is None when no port is specified.</comment>

<file context>
@@ -94,6 +94,38 @@ def fix_urljoin_bug(url: str, nesting_limit=5) -> str:
+        return url  # Child URL already has credentials
+
+    # Only inject if same host (hostname and port match)
+    if root_parsed.hostname != child_parsed.hostname or root_parsed.port != child_parsed.port:
+        return url
+
</file context>