Fixed a weakness in sniff where all whitespace on lines were stripped, in some cases eliminating desired blank columns.

lib/galaxy/datatypes/sniff.py

@@ -81,7 +81,7 @@ def get_headers(fname, sep, count=30):
     """
     headers = []
     for idx, line in enumerate(file(fname)):
-        line = line.strip()
+        line = line.rstrip('\n\r')
         if idx == count:
             break
         headers.append( line.split(sep) )
@@ -105,8 +105,11 @@ def is_column_based(fname, sep='\t'):
     False
     >>> is_column_based(fname, sep=' ')
     True
+    >>> fname = get_test_fname('test_ensembl.tab')
+    >>> is_column_based(fname)
+    True
     """
-    headers  = get_headers(fname, sep=sep)
+    headers = get_headers(fname, sep=sep)
     
     if not headers:
         return False

lib/galaxy/datatypes/test/test_ensembl.tab

@@ -0,0 +1,8 @@
+Ensembl Gene ID	Ensembl Transcript ID	Chicken Ensembl Gene ID	Orthology Type
+ENSOANG00000013952	ENSOANT00000021996	ENSGALG00000012404	ortholog_one2one
+ENSOANG00000013954	ENSOANT00000021998	ENSGALG00000012401	ortholog_one2one
+ENSOANG00000013954	ENSOANT00000021999	ENSGALG00000012401	ortholog_one2one
+ENSOANG00000013954	ENSOANT00000022000	ENSGALG00000012401	ortholog_one2one
+ENSOANG00000013955	ENSOANT00000022001		
+ENSOANG00000013955	ENSOANT00000022002		
+ENSOANG00000013956	ENSOANT00000022005	ENSGALG00000012400