lee/heritrix3
1
0
Fork 0
mirror of https://github.com/internetarchive/heritrix3.git synced 2026-09-23 14:15:47 +00:00
Code Issues Packages Projects Releases Wiki Activity

* ExtractorMultipleRegex.java

Browse Source 
fix omission from last commit, uriRegex
This commit is contained in:
Noah Levitt
2012-11-15 14:01:01 -08:00
parent 6ec4b5a973
commit 13e1f9d64d
1 changed files with 7 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/src/main/java/org/archive/modules/extractor/ExtractorMultipleRegex.java
+7 -2
View File
@@ -93,6 +93,11 @@ public class ExtractorMultipleRegex extends Extractor {
add(new GroupList(matcher.toMatchResult()));
}
}
public MatchList(GroupList... groupList) {
for (GroupList x: groupList) {
add(x);
}
}
};
protected class GroupList extends LinkedList<String> {
private static final long serialVersionUID = 1L;
@@ -111,8 +116,8 @@ public class ExtractorMultipleRegex extends Extractor {
// uri regex
Matcher matcher = TextUtils.getMatcher(getUriRegex(), curi.getURI());
if (matcher.matches()) {
matchLists = new LinkedHashMap<String, ExtractorMultipleRegex.MatchList>();
// XXX matchLists.put("uriRegex", );
matchLists = new LinkedHashMap<String,MatchList>();
matchLists.put("uriRegex", new MatchList(new GroupList(matcher)));
} else {
return; // if uri regex doesn't match, we're done
}