Hi Werner,
in first thanks for your answer.
I figure out some strange behaviour. When members invalidatedFrom,
invalidatedOn, invalidScale commented out (in mapping) an update works
fine. If uncomment fields in mapping --> results a ClassCastException
when processing field createdFrom
I taking three debug snapshoots in SQLEngine#toJava() line 471:
Case 1: comment out invalidatedFrom, invalidatedOn, invalidScale --> no
problem
object= BigDecimal = 1030
field= 0
column= 0
col=
convertFrom= SQLTypes$44 java.lang.Integer-->java.math.BigDecimal
convertTo= SQLTypes$13 java.math.BigDecimal-->java.lang.Integer
convertParam= null
name= "CREATED_FROM"
sqltype=2
Case 2: nothing comment out --> results ClassCastException
object= BigDecimal = 1030
field= 0
column= 0
col=
convertFrom= SQLTypes$44 java.lang.Integer-->java.math.BigDecimal
convertParam= null
convertTo= SQLTypes$13 java.math.BigDecimal-->java.lang.Integer
name= "CREATED_FROM"
sqlType= 2
Case 3: create new BradenScale object, object is in castor cache.
no step into.
As you can see above, un/comment some field in mapping leads to a
ClassCastException. I checked this for more then 10 time, to avoid any
false statement. Case 1 and case 2 IMHO are equals. This sounds strange,
but its the truth. I'm going slightly crazy.
>>Castor 0.9.6 seems to hard-cutted refactored in comparsion to 0.9.4.
>>Some minor negligences in JDO-config or code leads now in to big boom.
>>
>>
>Just to be sure here: Are you trying to say that we introduced too many errors
>into 0.9.6 by not paying due-diligence where
>we should have done so ?
>
>
No, i havent said: new version, new and more bug. My intent was to
critise the changing some default behaviours. E.g. castor 0.9.4 accept
field declarations without declaring java type. Castor 0.9.6 crashs, in
this case, with a log message 'unexpected NullPointerException' without
any hint. The fault for this problem was on my side, but its hard to
figure out where the failure located. Thats what i mean.
>>I cant find the simliar bug with MySQL 5.0 on codehouse issue page. Did
>>you know the bugnumber?
>>
>>
>There's no bug yet, as I am not quite sure where the problem is and whether
>it's reproducible. It could e.g. by down to the
>wrong jdbc driver being used when trying to access mySQL 5.0. Let me see what
>I can share with you in the next days.
>
>
>
Thanks, please let me know if you get further informations.
One question:
In wich code location will SQLEnginges for each class generated? If i
know about this location, i can step in to to recieve more information
of castor internals and type mapping?
Andre Teubner
