Fixed in Subversion. On Nov 10, 2010, at 2:23 PM, Andrus Adamchik wrote:
> Ok, I think this is closer: > > https://issues.apache.org/jira/browse/CAY-1505 > > I found this in my own application and am working on a fix now. > > Andrus > > > > On Nov 6, 2010, at 12:57 PM, Andrus Adamchik wrote: > >> Hi, >> >> Just found a bit unrelated issue with POST_LOAD callback: >> >> https://issues.apache.org/jira/browse/CAY-1503 >> >> For your case, it should help if you could post a code example with comments >> on where you are seeing and not seeing the callbacks. >> >> Andrus >> >> >> On Nov 1, 2010, at 8:10 PM, Bruno René Santos wrote: >> >>> Hello all, >>> >>> I am having a possibly strange behaviour on my lifecycle callbacks. Imagine >>> de >>> following scenario: >>> >>> 1 textfield and 1 table on a web application. I fill table with query >>> results >>> depending on the textfield value. Each time i get a valueChange event on the >>> textField I remove all lines from table and create new ones. As each line >>> is a >>> cayennedataobject, each time I do this procedure I call deleteObject for >>> each >>> old line that was removed when new lines are created. In the end I >>> commitChanges >>> and only the last lines are saved in relation with the superclass. >>> >>> The problem here is the following. For each erased line cayenne call the >>> preremove callback where I create a log for the table removal, in pending >>> state. >>> The state of the object is modified (this is the first weird thing, >>> shouldn’t be >>> new?) after the preRemove callback the desired process is done but Cayenne >>> never >>> calls the postRemove callback on this situation, for other regular deletes >>> everything works. >>> >>> Anybody recalls a situation where the preXX callback is called but not the >>> postXX? >>> >>> Thanx >>> Bruno Santos >>> >>> >> >> > >
