David E. Jones wrote:
On Feb 5, 2007, at 10:16 AM, Si Chen wrote:
David E. Jones wrote:
Hold on a minute there.... did you actually test and find this to be
a problem? The +/- notation is an entity engine ONLY thing and
should never make it to the database.
This patch should be reverted and if +/- are making it to the
database instead of being replaced with an ASC/DESC by the entity
engine then THAT bug should be fixed.
This is a slippery slope and we should backup to the top before it
gets going...
-David
On Feb 2, 2007, at 6:16 PM, [EMAIL PROTECTED] wrote:
Author: sichen
Date: Fri Feb 2 17:16:36 2007
New Revision: 502824
URL: http://svn.apache.org/viewvc?view=rev&rev=502824
Log:
Fix a pretty significant bug with sequence of inventory item
reservations for orders: FIFO and LIFO were reversed (I checked
this) and the +/- notation does not work well with PostgreSQL
Modified:
ofbiz/trunk/applications/product/script/org/ofbiz/product/inventory/InventoryReserveServices.xml
Modified:
ofbiz/trunk/applications/product/script/org/ofbiz/product/inventory/InventoryReserveServices.xml
David,
I worked on this with Leon, and we did test this a few times, and the
inventory reservation sequence was wrong. Previously when using the
FIFO reservation, the last item received was being reserved against
orders. It seems that "-datetimeReceived" was ordering them by
descending order of date time received, and as a result reservations
were done in the wrong sequence.
We did not realize that the + or - notation were an entity engine
thing, but please test the actual reservation of inventory before and
after this patch and let us know which behavior you feel is correct.
I realize this is a pretty basic fix to an existing feature, and I
had trouble believing that something like this could be broken, but
we did try it several times ourselves.
Si
Perhaps I should have been more explicit in my comment. I was
referring only to the apparent bug report about the +/- notation on
certain databases, including Postgres.
My comments did not mention nor had anything to do with the fix of the
inventory reservation sequence, and it's great that you guys found and
fixed that bug. It looks like the _REC reserve orders were added to
the original code but not implemented correctly, so yeah, it's great
that you took care of that.
-David
Ok. Actually I do not think there is a problem with +/- notation. I
was trying to figure out what was wrong with the inventory reservation
and realized that it was not a postgresql notation, so I made a comment
about it on my commit, but I don't have any reason to believe it's a
source of problem.