Mike- Thanks for looking into this problem so carefully. If you want to create a bug report, you can do so at
http://jira.codehaus.org/browse/CASTOR If you, please include your proposed fix as it will be a great help in starting the investigation on our end. But your solution sounds like a good one, and it should be a pretty easy fix. Thanks, Stephen On 4/4/06, Mike Gozaloff <[EMAIL PROTECTED]> wrote: > > Thomas, > > I ran across the same problem today. After debugging through the source, it > appears to be a logic problem in SchemaWriter. In the (many) places where > it checks the list of cached included schemas to see if a type is defined in > an include, the boolean field "found" is not handled correctly. It retains > the value from the last included schema - not all the included schemas as it > should. This makes the writer define (inline) all types in all included > schemas except the last schema in the _cachedincludedSchemas Hashtable. > > A simple change from this: > while (includedSchemas.hasMoreElements()) > to this: > while (includedSchemas.hasMoreElements() && !found) > > in all the places the included schemas are checked, should kick it out of > the loop (retaining the value of found) and fix the problem. > > I have not tested this fix nor figured out how to submit a Castor bug yet. > > FWIW, > > gozer > > > -- > View this message in context: > http://www.nabble.com/problem-while-doing-Schema.cacheIncludedSchema%28%29-twice-t1333266.html#a3757332 > Sent from the Castor - User forum at Nabble.com. > > > ------------------------------------------------- > If you wish to unsubscribe from this list, please > send an empty message to the following address: > > [EMAIL PROTECTED] > ------------------------------------------------- > >
