171639 - dupes not marked in original bug

Status: RESOLVED FIXED

(Bugzilla :: Creating/Changing Bugs, defect)

Description

[David Avery]

since the groups checkin marking a bug a dupe adds the comment to the dupe 
but nothing shows in the original bug

Comment 1

[David Avery]

research shows that process bug is setting the isprivate bit unconditionally 
on the dupe comment in the original bug 

            SendSQL("INSERT INTO cc (who, bug_id) VALUES ($reporter, " .
SqlQuote($duplicate) . ")");
        }
======== bad line ======
        AppendComment($duplicate, $::COOKIE{'Bugzilla_login'}, "*** Bug
$::FORM{'id'} has been marked as a duplicate of this bug. ***", 1);
=======================
        CheckFormFieldDefined(\%::FORM,'comment');
        SendSQL("INSERT INTO duplicates VALUES ($duplicate, $::FORM{'id'})");


--------------------

confirmed on landfill [00:10:40] <bbaetz> also:
<bbaetz> [% isinsider = Param("insidergroup") &&
UserInGroup(Param("insidergroup")) %]
<justdave> I could swear I pointed that out to someone in a review comment a
while back...
<bbaetz> so if there isn't an insider group, isinsider is false, and so you
never see it
<justdave> if there is no insider group everything should be visible, right?
<bbaetz> well, what if you deprecate the insider group?



also in my case the "insidersgroup" param was lowercase "insider" while 
the group name was "Insider", which had worked up till now because mysql
does case insensitive string compares

Comment 2

[Bradley Baetz (:bbaetz)]

The reason that the case started to matter is because of the hash cache, which
means that we now compare using perl, instead of mysql. I don't see a problem
with this, except we should document it in the releae notes somewhere.

Joel, you may want to use lc() to compare stuff in your product group rewrite,
though.

Comment 3

[Matthew Tuck [:CodeMachine]]

Can't we just fix this up?

AIUI string compares should be going away eventually, as the product group and
the system role rewrites would be using IDs rather than names.

Comment 4

[Bradley Baetz (:bbaetz)]

no, we're still going to be using strings for comparisions like 'editusers'.
Even if you have a 'role' instead of a 'group', you'd still compare on that.

Yes, the product group rewrite will reduce this, but if the param is a string,
then theres not much more we can do....

Comment 5

[Matthew Tuck [:CodeMachine]]

That's not my intention.  The new implementation needs to properly handle
renaming like everything else.

It might be displayed as a string, but should be stored as an ID, and in that
sense there probably will be string lookups I suppose.

Comment 6

[Joel Peshkin]

So, before I change anything on this....

1) Are we saying that all group name comparisons should be done in a
case-insensitive manner?

2) For installations with private comments, do we want to have dupes comments
private or not?  (Since the original discussion on this, it has become possible
to go back and adjust a privacy flag, so that could alter the decision)

Comment 7

[Bradley Baetz (:bbaetz)]

1) I say no (with the proviso that your other patch for product groups should do
a case-insensitive compare, just for sanity checks)

2) Can;t you use the $::FORM value? Ie make the dupe comment in the other bug
private if the comment you're currently making is private?

Comment 8

[Joel Peshkin]

Attachment: Changes default privacy on duplication comment

Actually, the right thing to do on this now is to let the duplication comment
be public by default.  None of the information from the comment on the dupe is
included in the other bug anyway and, more importantly, a use without access to
the  dupe will not see anything bad (like the title) in the bug link.

Comment 9

[Bradley Baetz (:bbaetz)]

Comment on attachment 101100 [details] [diff] [review]
Changes default privacy on duplication comment

I'm happy with this if you are. Please do add some sort of comment, though.

Comment 10

[Joel Peshkin]

Checking in process_bug.cgi;
/cvsroot/mozilla/webtools/bugzilla/process_bug.cgi,v  <--  process_bug.cgi
new revision: 1.151; previous revision: 1.150
done