diff options
| author | Bryan Newbold <bnewbold@archive.org> | 2020-09-02 10:23:49 -0700 |
|---|---|---|
| committer | Bryan Newbold <bnewbold@archive.org> | 2020-09-02 10:23:49 -0700 |
| commit | 6e13949bf0d602c792643bdfcc1acafc08fb07cd (patch) | |
| tree | b3bd081ad66ccb6e576427df223c73989a479bc8 | |
| parent | e76eab9a1b796b436ec2cb7dc7ad8ea00b0580ee (diff) | |
| download | sandcrawler-6e13949bf0d602c792643bdfcc1acafc08fb07cd.tar.gz sandcrawler-6e13949bf0d602c792643bdfcc1acafc08fb07cd.zip | |
more bad SHA1 PDF
| -rw-r--r-- | python/sandcrawler/pdfextract.py | 2 |
1 files changed, 2 insertions, 0 deletions
diff --git a/python/sandcrawler/pdfextract.py b/python/sandcrawler/pdfextract.py index da6b9f6..06868a7 100644 --- a/python/sandcrawler/pdfextract.py +++ b/python/sandcrawler/pdfextract.py @@ -37,6 +37,7 @@ BAD_PDF_SHA1HEX = [ "447fa6b5a90742a86429a932f6608d8e141688c0", "4c81129904f7976a50825595a3497ea7b52579ef", "50b3c5a3122272aca69855ef06b85d0b43a76eb1", + "58d9ae7dcb0a7dbbdfc58ad266030b037e9cd0ff", "5e6a3adde9f08c276c4efd72bfacb256f2ec35d9", "646c4a654270606256397684204ff0f3d17be2e7", "64d821d728f9a3dc944b4c03be00feea0b57e314", @@ -51,6 +52,7 @@ BAD_PDF_SHA1HEX = [ "b2b66b9c7f817a20144456f99c0be805602e8597", "b2d719120306b90eb8dd3580b699a61ec70556f4", "b5bf8b7467fb095c90adf3b49aa1687291e4469c", + "c1b583fbd052572f08158d39ffe4d7510dadbebb", "ccb1debcfae006a3fc984e9e91309b9706a5c375", "cd8a7c3b8d850ebedc1ca791ccb37b9a2689f9c3", "d17b1e254cce82df5c6eb4fd492cef91e7e11558", |