diff options
Diffstat (limited to 'webmentions_ssg/tasks')
| -rw-r--r-- | webmentions_ssg/tasks/scanner.py | 15 |
1 files changed, 11 insertions, 4 deletions
diff --git a/webmentions_ssg/tasks/scanner.py b/webmentions_ssg/tasks/scanner.py index e45085d..666eba7 100644 --- a/webmentions_ssg/tasks/scanner.py +++ b/webmentions_ssg/tasks/scanner.py @@ -25,6 +25,7 @@ from sqlalchemy.orm import selectinload from .. import DATABASE as db from .. import HUEY as huey +from ..config import normalize_ignored_hostnames from ..models import SentWebmention, Source from ..url_security import is_http_url from .sender import send_webmention @@ -348,9 +349,16 @@ def scan_source_file( Webmention source. :raises OSError: If the source document cannot be read. """ - document = BeautifulSoup(path.read_bytes(), "html5lib") + # Use the path as published for its URL, but read only files whose + # resolved target stays inside the configured source directory. Resolving + # before reading also protects against symlinked index.html files. + resolved_root = root.resolve() + resolved_path = path.resolve() + if not resolved_path.is_relative_to(resolved_root): + raise SourceScanError(f"Source document escapes source directory: {path}") relative_path = path.relative_to(root) + document = BeautifulSoup(resolved_path.read_bytes(), "html5lib") source_url = canonical_url(document, relative_path=relative_path, base_url=base_url) @@ -515,9 +523,8 @@ def scan_sources_task() -> None: raise RuntimeError("SOURCE_BASE_URL must be an absolute HTTP or HTTPS URL") base_url = f"{base_url.rstrip('/')}/" - ignored_hostnames = tuple( - pattern.lower().rstrip(".") - for pattern in current_app.config.get("IGNORED_HOSTNAMES", ()) + ignored_hostnames = normalize_ignored_hostnames( + current_app.config["IGNORED_HOSTNAMES"] ) sources_by_path = { |
