Fix robustness across gitdm tools: NameError, LogPatchSplitter call, CSV column shape, pickling modes, and guarded pdb

src/cncfdm.py

@@ -688,10 +688,11 @@ def is_svntag(logpatch):
         AddDateLines (pa.date, max (pa.added, pa.removed))
         empl = pa.author.emailemployer (pa.email, pa.date)
         if not empl:
-            # Usually missing the final affiliation (like last affiliation ws untin 2019-01-01 and this patch is from 2019-03-03
-            print 'pdb on email ', pa.email
-            print 'pdb on logpatch ', logpatch
-            pdb.set_trace()
+            # Usually missing the final affiliation (like last affiliation was until some date and this patch is after that)
+            if DebugHalt:
+                pdb.set_trace()
+            else:
+                sys.stderr.write('Missing affiliation for %s at %s; skipping.\n' % (pa.email, pa.date))
             continue
         empl.AddCSet (pa)
         for sobemail, sobber in pa.sobs:

src/committags.py

@@ -43,6 +43,6 @@
             DB[m.group(1)] = Tag
 
 print 'Found %d commits, %d tags' % (len(DB.keys()), Tags)
-out = open('committags.db', 'w')
+out = open('committags.db', 'wb')
 pickle.dump(DB, out)
 out.close()

src/database.py

@@ -188,7 +188,7 @@ def AllAffsCSV(file, hlist):
                 if em in emails:
                     print 'This is bad, reverse email already in emails, check: `em`, `email`, `emails`'
                     pdb.set_trace()
-                writer.writerow ([email_encode(em), email_encode(name), emplstr, datestr])
+                writer.writerow ([email_encode(em), email_encode(name), emplstr, datestr, source])
 
 def AllHackers ():
     return HackersByID.values ()

src/gitdm.py

@@ -413,7 +413,8 @@ def TrimLTSOBs (p):
 #
 print >> sys.stderr, 'Grabbing changesets...\r',
 
-patches = logparser.LogPatchSplitter(sys.stdin)
+# Default to a wide-open date range to match cncfdm.py
+patches = logparser.LogPatchSplitter(sys.stdin, datetime.datetime(1970, 1, 1), datetime.datetime(2069, 1, 1))
 printcount = CSCount = 0
 
 for logpatch in patches:

src/linetags.py

@@ -66,7 +66,7 @@ def MapCommits():
 #
 # Grab the tags/version database.
 #
-dbf = open('committags.db', 'r')
+dbf = open('committags.db', 'rb')
 DB = pickle.load(dbf)
 dbf.close()