better must-match pattern for intranet file-crawls

git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@7218 6c8d7289-2bf4-0310-a012-ef5d649a1542
yacy · Oct 4, 2010 · 574346f · 574346f
1 parent aacf572
commit 574346f
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/htroot/Crawler_p.java b/htroot/Crawler_p.java
@@ -144,7 +144,7 @@ public static serverObjects respond(final RequestHeader header, final serverObje
                 if (newcrawlingMustMatch.length() < 2) newcrawlingMustMatch = CrawlProfile.MATCH_ALL; // avoid that all urls are filtered out if bad value was submitted
                 // special cases:
                 if (crawlingStartURL!= null && fullDomain) {
-                    newcrawlingMustMatch = crawlingStartURL.isFile() ? "file:///.*" : crawlingStartURL.isSMB() ? "smb://.*" : ".*" + crawlingStartURL.getHost() + ".*";
+                    newcrawlingMustMatch = crawlingStartURL.isFile() ? "file://" + crawlingStartURL.getPath() + ".*" : crawlingStartURL.isSMB() ? "smb://" + crawlingStartURL.getPath() + ".*" : ".*" + crawlingStartURL.getHost() + ".*";
                 }
                 if (crawlingStart!= null && subPath && (pos = crawlingStart.lastIndexOf('/')) > 0) {
                     newcrawlingMustMatch = crawlingStart.substring(0, pos + 1) + ".*";