diff options
| author | Bryan Newbold <bnewbold@archive.org> | 2022-12-16 11:17:24 -0800 |
|---|---|---|
| committer | Bryan Newbold <bnewbold@archive.org> | 2022-12-16 11:17:24 -0800 |
| commit | b42103eb35f8eb55bda03facb8b14a366fd544c2 (patch) | |
| tree | 587ea854e5f0707c5f1393171be345a5bfaaf418 | |
| parent | 5b1d9a31e5f0bfd0e2d544d87953043ac8f8f8b9 (diff) | |
| download | sandcrawler-b42103eb35f8eb55bda03facb8b14a366fd544c2.tar.gz sandcrawler-b42103eb35f8eb55bda03facb8b14a366fd544c2.zip | |
bad pdf hash
| -rw-r--r-- | python/sandcrawler/pdfextract.py | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/python/sandcrawler/pdfextract.py b/python/sandcrawler/pdfextract.py index 5f2dc1a..97d338e 100644 --- a/python/sandcrawler/pdfextract.py +++ b/python/sandcrawler/pdfextract.py @@ -144,6 +144,7 @@ BAD_PDF_SHA1HEX: List[str] = [ "d188762a7e3ab5d4ee8a897204316513e4e636ec", "d613b9e4442f5d5d19ea6814fa9729bff7da7c85", "d6b0f405bf13c23d0e90c54eea527442786d1cd3", + "d91d3830bf455e6dd782eee46218e35d29f07dfd", "da2211ee2dbc6dda36571976d810e2366a3d2504", "dbb3093a797e0ae83d39eb7b235ff85a17fd965c", "e01bb7256d77aea258313bb410dfcfc10512f420", |