[Building Sakai] SAM-2185 and SAM-2241

[Bryan Holladay]

It sounds like the same issue as SAM-2185 and, as Kevin mentions
in SAM-2241, they aren't connected (just two of the same errors in
different places). All I did in SAM-2185 was to make sure the ID was null
when it was a new item (not sure how or where that ID was being set). I
looked through the code for SAM-2241 briefly and didn't see anything
obvious. Without spending hours and tracing through the code, I can't be
much help.

-Bryan

On Wed, Mar 4, 2015 at 5:04 PM, George Pipkin <gpp8p at virginia.edu> wrote:

>  Hi Bryan -
>
>
>    This damn thing has landed in my lap, and I've spent a day on it
> getting nowhere.  This is the one where if the student checks
> "marked for review" on an MC, it bombs way down in Hibernate because it
> tries to insert a duplicate ASSESSMENTGRADINGID into
> SAM_ITEMGRADING_T.  I've done a lot of tracing, and I looked at your
> patch  (SAM-2185.patch )which doesn't kill this particular error.
> In your comments, you said you thought the problem revolved around
> itemGradingId, which is nulled out by your patch.  I was wondering
> if you could expand a little on that.  I'm kind of lost, and I don't
> understand what's going on.  I have noticed one difference between what
> happens when the mark for review boxes are checked (and it throws the
> exception) and when they are not (and it works fine):
>
> Your patch introduces the following line at line 587
> SubmitToGradingActionListener:
>
>             } else { // itemGrading from new set doesn't exist, add to set
> in
>                 // this case
>                 // log.debug("**** SubmitToGrading: need add new item");
>                     //a new item should always have the grading ID set to
> null
>  ---->                   newItem.setItemGradingId(null);
>                 newItem.setAgentId(adata.getAgentId());
>                 updateItemGradingSet.add(newItem);
>             }
>
> In the case where the mark for review boxes are not checked, this part of
> the code is traversed ONE time
>
> In the case where the boxes are checked, it is traversed twice, and right
> before the save that generates the
> exception, it is traversed four times.
>
> I was wondering if you might have some idea if this is significant or not.
>
>      - George Pipkin
>         U.Va.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://collab.sakaiproject.org/pipermail/sakai-dev/attachments/20150305/e6943537/attachment.html