[Rt-devel] Bug in Query builder (3.4.5 and 3.6.1)?

[Philip Kime]

I did a quick test of 3.6.2rc1 and this bug appears not to be fixed in
that release. Will the fix make it to 3.6.2 RTM? Michael's fix, fixed
the problem if applied to 3.6.1, by the way.

PK

-----Original Message-----
From: Jesse Vincent [mailto:jesse at bestpractical.com] 
Sent: Wednesday, November 08, 2006 11:41 AM
To: Philip Kime
Cc: Michael Gilbert; Jesse Vincent; rt-devel at lists.bestpractical.com
Subject: Re: [Rt-devel] Bug in Query builder (3.4.5 and 3.6.1)?




On Wed, Nov 08, 2006 at 11:35:39AM -0800, Philip Kime wrote:
> I haven't installed 3.4.6 or 3.6.2rc1 yet but I checked the code in 
> their Build.html and this patch hasn't been applied to either.

Sometimes, there are multiple patches that can fix the same issue. I
can't 100% assert that that would be the case here, but a quick test
would be much appreciated.


> PK
> 
> 
> 
> -----Original Message-----
> From: Michael Gilbert [mailto:mdgilb at u.washington.edu]
> Sent: Monday, November 06, 2006 12:30 PM
> To: Jesse Vincent
> Cc: Philip Kime; rt-devel at lists.bestpractical.com
> Subject: Re: [Rt-devel] Bug in Query builder (3.4.5 and 3.6.1)?
> 
> I believe we had this same problem a while back, and the solution 
> suggested at 
> http://lists.bestpractical.com/pipermail/rt-devel/2005-August/007507.h
> tm
> l
> seemed to work fine for us (This may be in 3.6.2rc1, but if an upgrade

> isn't desirable the patch is below).
> 
> --- html/Search/Build.html      Thu Aug 25 17:40:12 2005
> +++ html/Search/Build.html      Mon Aug 29 21:31:15 2005
> @@ -595,6 +595,7 @@
>          }
>          elsif ( $current & AGGREG ) {
>              $ea   = $val;
> +            $parentnode->setNodeValue($ea);
>              $want = KEYWORD | PAREN;
>          }
>          elsif ( $current & KEYWORD ) { @@ -652,7 +653,6 @@
> 
>              # explicity add a child to it
>              $lastnode = RT::Interface::Web::QueryBuilder::Tree->new(
> $clause, $parentnode );
> -            $lastnode->getParent()->setNodeValue($ea);
> 
>              ( $ea, $key, $op, $value ) = ( "", "", "", "" );
>          }
> 
> -michael
> 
> Jesse Vincent wrote:
> 
> >Can you test out 3.4.6 and 3.6.2rc1?  I believe we dealt with this.
> >
> >
> >On Thu, Nov 02, 2006 at 07:57:18PM -0800, Philip Kime wrote:
> >  
> >
> >>Hmm, but what about this - in the advanced view, I do
> >>
> >>Queue = 'A' OR ( Status = 'new' AND Subject LIKE 'office')
> >>
> >>I click apply and it becomes:
> >>
> >>Queue = 'A' AND ( Status = 'new' AND Subject LIKE 'office')
> >>
> >>??
> >> 
> >>
> >>-----Original Message-----
> >>From: Joby Walker [mailto:joby at u.washington.edu]
> >>Sent: Thursday, November 02, 2006 7:26 PM
> >>To: Philip Kime
> >>Cc: rt-devel at lists.bestpractical.com
> >>Subject: Re: [Rt-devel] Bug in Query builder (3.4.5 and 3.6.1)?
> >>
> >>This isn't broken, you need to use parenthesis:
> >>
> >>A,
> >>OR (
> >>    B,
> >>    AND C
> >>    )
> >>
> >>jbw
> >>
> >>Philip Kime wrote:
> >>    
> >>
> >>>Can anyone reproduce this?
> >>> 
> >>>Construct a query "A AND B" then change the aggregator to "OR" and 
> >>>add
> >>>      
> >>>
> >>>a condition "C.
> >>> 
> >>>The original AND changes to an OR. so it becomes "A OR B OR C"
> >>> 
> >>>Same thing happens if you start with OR and add an AND, you end up 
> >>>with "A AND B AND C".
> >>> 
> >>>Happens in 3.4.5 and 3.6.1
> >>> 
> >>>The same thing constructed on the command-line with the REST 
> >>>interface
> >>>      
> >>>
> >>>isn't broken and returns the right thing.
> >>> 
> >>>PK
> >>> 
> >>>--
> >>>Philip Kime
> >>>NOPS Systems Architect
> >>>310 401 0407
> >>> 
> >>>
> >>>
> >>>-------------------------------------------------------------------
> >>>--
> >>>-
> >>>--
> >>>
> >>>_______________________________________________
> >>>List info: 
> >>>http://lists.bestpractical.com/cgi-bin/mailman/listinfo/rt-devel
> >>>      
> >>>
> >>_______________________________________________
> >>List info: 
> >>http://lists.bestpractical.com/cgi-bin/mailman/listinfo/rt-devel
> >>
> >>    
> >>
> >
> >  
> >
> 

--