correctly handle comment assignment

* handles the ***'s literal case
* handles the private comment case
* handles the stripping of legacy attachment notice

2 files changed, 24 insertions, 14 deletions

diff --git a/bugzilla_update_user_comments.py b/bugzilla_update_user_comments.py
index b3009e6..3acc4c5 100755
--- a/bugzilla_update_user_comments.py
+++ b/bugzilla_update_user_comments.py
@@ -54,22 +54,15 @@ def update(user):
         jcom, jxcom = comdetails[0]
         coms = json.loads(jcom)
         xcoms = json.loads(jxcom)
-
         for key, xi in xcoms.iteritems():
             com = coms[util.get_index(coms, "count", int(key))]
             content = com['text']
-
-            if com['is_private']:
-                vlog('is private')
-                comment_content = bzlib.security_mask
-            else:
-                vlog('is NOT private')
-                comment_content = content + xi['xattached']
-
+            comment_content = bzlib.build_comment_comment_regen(content, com['is_private'])
             if com["creator"] == user['user']:
                 log("Updating comment %s for %s" % (xi['xctransaction'], user['user']))
                 phabdb.set_comment_author(xi['xctransaction'], user['userphid'])
-                phabdb.set_comment_content(xi['xctransaction'], comment_content)
+                phabdb.set_comment_content(xi['xctransaction'],
+                                           comment_content + xi['xattached'])
 
     current = phabdb.get_user_migration_comment_history(user['user'], pmig)
     if current:
diff --git a/wmfphablib/bzlib.py b/wmfphablib/bzlib.py
index 92fc7ef..600c79e 100755
--- a/wmfphablib/bzlib.py
+++ b/wmfphablib/bzlib.py
@@ -2,6 +2,10 @@ import re
 
 prepend = 'bz'
 security_mask = '//**content hidden as private in Bugzilla**//'
+dupe_literals = ['This bug has been marked as a duplicate of bug',
+                 'This bug has been marked as a duplicate of',
+                 'has been marked as a duplicate of this bug']
+
 
 # Some issues are just missing instead of constant failures we skip
 missing = [15368, 15369, 15370, 15371, 15372, 15373, 15374]
@@ -28,10 +32,6 @@ def build_comment(c, secstate):
 
     # these indicate textual metadata history and should be
     # preserved as literal
-    dupe_literals = ['This bug has been marked as a duplicate of bug',
-                     'This bug has been marked as a duplicate of',
-                     'has been marked as a duplicate of this bug']
-
     clean_c = {}
     clean_c['author'] =  c['author'].split('@')[0]
     clean_c['creation_time'] = str(c['creation_time'])
@@ -63,6 +63,23 @@ def build_comment(c, secstate):
     clean_c['text'] = c['text']
     return clean_c
 
+def build_comment_comment_regen(text, is_private):
+
+    if is_private:
+        text = security_mask
+
+    fmt_text = []
+    text = text.splitlines()
+    for t in text:
+        if t.startswith('Created attachment'):
+            continue
+        elif '***' in t and any(map(lambda l: l in t, dupe_literals)):
+            fmt_text.append('%%%{0}%%%'.format(t))
+        else:               
+            fmt_text.append(t)
+    clean_text = '\n'.join(fmt_text)
+    return clean_text
+
 def find_attachment_in_comment(text):
     """Find attachment id in bz comment
     :param text: str