diff options
author | Andreas Baumann <abaumann@yahoo.com> | 2014-07-25 09:28:51 +0200 |
---|---|---|
committer | Andreas Baumann <abaumann@yahoo.com> | 2014-07-25 09:28:51 +0200 |
commit | e53f13ccaa3a0b595351fc536bf1a6d6be805175 (patch) | |
tree | ecfb7df20e09841283776ca928fbe49f2428a22b | |
parent | aea482c8b49e9c4fb5e07ee84d80106cdada9f96 (diff) | |
download | crawler-e53f13ccaa3a0b595351fc536bf1a6d6be805175.tar.gz crawler-e53f13ccaa3a0b595351fc536bf1a6d6be805175.tar.bz2 |
we also have to normilze with query
-rwxr-xr-x | src/modules/urlnormalizer/googleurl/GoogleURLNormalizer.cpp | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/src/modules/urlnormalizer/googleurl/GoogleURLNormalizer.cpp b/src/modules/urlnormalizer/googleurl/GoogleURLNormalizer.cpp index b09b1f3..c8e50a6 100755 --- a/src/modules/urlnormalizer/googleurl/GoogleURLNormalizer.cpp +++ b/src/modules/urlnormalizer/googleurl/GoogleURLNormalizer.cpp @@ -102,7 +102,8 @@ URL GoogleURLNormalizer::normalize( const URL url, const string s ) componentString( canonical, parsed.host ), port, componentString( canonical, parsed.path ), - "", "" ); + componentString( canonical, parsed.query ), + "" ); } REGISTER_MODULE( "google_urlnormalizer", URLNormalizer, GoogleURLNormalizer ) |