X-Git-Url: https://projects.mako.cc/source/wikiq/blobdiff_plain/c695f50b7f88b97d12a827338a6dfd2c53180c17..d470ebe8c38c0df66a84db23ef32681278bda2fe:/wikiq.cpp

diff --git a/wikiq.cpp b/wikiq.cpp
index 18259ea..3451582 100644
--- a/wikiq.cpp
+++ b/wikiq.cpp
@@ -11,7 +11,6 @@
 #include <stdlib.h>
 #include "expat.h"
 #include <getopt.h>
-#include "disorder.h"
 #include "md5.h"
 #include "dtl/dtl.hpp"
 #include <vector>
@@ -345,7 +344,6 @@ write_row(revisionData *data)
         << data->editorid << "\t"
         << ((data->minor) ? "TRUE" : "FALSE") << "\t"
         << (unsigned int) data->text_size << "\t"
-        << shannon_H(data->text, data->text_size) << "\t"
         << md5_hex_output << "\t"
         << reverted_to << "\t"
         << (int) additions.size() << "\t"
@@ -540,14 +538,15 @@ void print_usage(char* argv[]) {
          << "a tab-separated stream of revisions on standard out:" << endl
          << endl
          << "title, articleid, revid, timestamp, anon, editor, editorid, minor," << endl
-         << "text_length, text_entropy, text_md5, reversion, additions_size, deletions_size" << endl
+         << "text_length, text_md5, reversion, additions_size, deletions_size" << endl
          << ".... and additional fields for each regex executed against add/delete diffs" << endl
          << endl
          << "Boolean fields are TRUE/FALSE except in the case of reversion, which is blank" << endl
          << "unless the article is a revert to a previous revision, in which case, it" << endl
          << "contains the revision ID of the revision which was reverted to." << endl
          << endl
-         << "author: Erik Garrison <erik@hypervolu.me>" << endl;
+         << "authors: Erik Garrison <erik@hypervolu.me>" << endl
+         << "         Benjamin Mako Hill <mako@atdot.cc>" << endl;
 }
 
 
@@ -635,14 +634,13 @@ main(int argc, char *argv[])
     cout << "title" << "\t"
         << "articleid" << "\t"
         << "revid" << "\t"
-        << "date" << " "
+        << "date" << "_"
         << "time" << "\t"
         << "anon" << "\t"
         << "editor" << "\t"
         << "editor_id" << "\t"
         << "minor" << "\t"
         << "text_size" << "\t"
-        << "text_entropy" << "\t"
         << "text_md5" << "\t"
         << "reversion" << "\t"
         << "additions_size" << "\t"