Hi Kristis,
On Sun, Jun 29, 2008 at 5:10 PM, Kristis Makris <[EMAIL PROTECTED]> wrote:
> Hi Javor
>
> On Sun, 2008-06-29 at 14:21 +0300, Javor Nikolov wrote:
>> not the case of the resolution status which matters, it's the "status"
>> keyword itself. resolution_status_case_sensitive_verification => 0
>> should also apply to it, doesn't it?
>
> As Oliver mentioned, it doesn't.
OK, now I understand that "Apply a case sensitive resolution and
resolution status verification" does not apply to the "status" keyword
itself which should be always lower-case (whit the default regex
expressions in glue.conf).
>
>> 2) Mixed-case status
>> Status 6: RESOLVED FIXED
>> This text "Status 6: RESOLVED FIXED" remains part of bugzilla comment
>> body and the bug status is not changed. However, as I mentioned before
>> - the same expression is anyway recognized as attempt to change bug
>> status since when applied to an already fixed bug (or new one) -
>
> I doubt this is the case. Because Integration.pm:process_activity_commit has
> been written to first apply the comment and then attempt to change the status.
> Are you *sure* the bug was not FIXED when you tried to commit with this log
> message ?
>
Sorry I haven't been clear enough. The point is that looks like a
mixed-case "Status" keyword, sometimes means an attempt for status
change and sometimes not:
A) When current bug status is ASSIGNED and the status change (to
RESOLVED FIXED) is a valid one - then a new comment appears in
bugzilla for that bug and the text "Status 6: RESOLVED FIXED" just
remains in the body of that comment; and bug status is unchanged. OK -
Status is not all-lower-case so as it shouldn't be processed like
status change attempt at all (it's just a part of svn commit log
message).
B) In the same conditions like (A) but with with lower-case "status"
keyword in the svn log message ("status 6: RESOLVED FIXED"). Then a
new comment is entered in where the "status 6: RESOLVED FIXED" is
striped-off and bug status is changed to RESOLVED FIXED.
OK - a valid status change attempt recognized and processed as expected.
C) Current status is NEW and a mixed-case "Status" message (Status 7:
RESOLVED FIXED). Then I get error message:
** Scmbug error 7: Bug '7' is in the 'NEW' state, instead of the
'ASSIGNED' or 'REOPENED' states.
(and as a result the commit to svn fails, and no changes in bugzilla either)
Why? I agree that transition from NEW to RESOLVED FIXED is not a valid
one but we don't have an all-lower-case "status" here. It shouldn't be
processed like status change at all. Shouldn't it be processed like in
A as a regular svn log text - i.e. "Status: ..." to remain just a part
of bugzilla comment log.
It's true that this error message in that case would be actually a
good thing in most of the cases since the mixed-case "Status" would be
a spelling mistake. Then why don't we also have such prevention in
case (A)? Maybe it makes sense such case-spell-checking to be
optional.
D) All like in C but current bug status is already RESOLVED FIXED. The
behaviour is the same as in C.
>> (Why do I get the last error message printed twice also seems strange).
>
> That's a known bug that should be fixed in the next release.
>
>> Seems that presence_of_bug_ids doesn't make a difference - I've tested
>> the above both with required and optional. All the other glue.conf
>
> What do you mean by "presence_of_bug_ids doesn't make a difference" ? What was
> your complete log message when you tried both with this policy set to required
> and optional ? What you have described in (1) and (2) so far does not indicate
> a log comment.
Well, I mentioned that since you suggested that maybe we're hitting a
variation of bug http://bugzilla.mkgnu.net/show_bug.cgi?id=1124 (which
is "Issues with status resolution when presence_of_bug_ids is
optional"). I meant that it didn't seem to influence the the behaviour
which I described /with "Status" keyword case/. (Not that required
doesn't prevent commits without bug id).
The complete log message (in case C above) is like that:
bug 7:
Scm integration test - try again to resolve new bug.
Status 7: RESOLVED FIXED
_______________________________________________
scmbug-users mailing list
[email protected]
http://lists.mkgnu.net/cgi-bin/mailman/listinfo/scmbug-users
