diff options
| author | Bryan Newbold <bnewbold@archive.org> | 2020-08-08 13:29:19 -0700 |
|---|---|---|
| committer | Bryan Newbold <bnewbold@archive.org> | 2020-08-08 13:29:19 -0700 |
| commit | 0aa723392c1c72a354731aa21c06c55adeacab30 (patch) | |
| tree | 2e1d938b4ada77ae9c9c6e5f44598f71814af7d6 | |
| parent | ae9e876db3f53f0d56521a69e6bc0104d607e783 (diff) | |
| download | sandcrawler-0aa723392c1c72a354731aa21c06c55adeacab30.tar.gz sandcrawler-0aa723392c1c72a354731aa21c06c55adeacab30.zip | |
another PDF hash to skip
| -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 29affa8..65df25d 100644 --- a/python/sandcrawler/pdfextract.py +++ b/python/sandcrawler/pdfextract.py @@ -35,6 +35,7 @@ BAD_PDF_SHA1HEX = [ "447fa6b5a90742a86429a932f6608d8e141688c0", "4c81129904f7976a50825595a3497ea7b52579ef", "50b3c5a3122272aca69855ef06b85d0b43a76eb1", + "5e6a3adde9f08c276c4efd72bfacb256f2ec35d9", "646c4a654270606256397684204ff0f3d17be2e7", "64d821d728f9a3dc944b4c03be00feea0b57e314", "88edcbab1cac2d70af5870422974afc253f4f0c6", |