Kai, I've worked with generated source in production projects (even if not so much with castor) and I would think there are two possible responses:
1) Use the source generator as a way to get started. Then maintain the schema/dtd+castormapping+javaobjects by hand thereafter, no longer utilizing the source generator. 2) Do not put any functionality aside from what the source generator generates in the castor objects. Use them only for persistence so that any bug fixes would be structural changes that would require updating the schema/dtd and then source regeneration. Either way, you just don't want to be in a position where you ever regenerate your source and then have to touch anything by hand. That's when it starts getting expensive. Personally, since #2 never seems to work right, #1 is where I usually land. Hth, ken. -----Original Message----- From: Mai Kai Zhen [mailto:[EMAIL PROTECTED]] Sent: Monday, January 06, 2003 5:42 AM To: [EMAIL PROTECTED] Subject: [castor-dev] how high is the cost of maintenence of castor xml in real world? This is a rather general question. I can see how java xml binding can save me time up front. I am wondering about the maintenence cost. People generate a set of classess from either schema or dtd to do marshall and unmarshall. Should the generated source files be put in source control? If there's a bug in the generated file, it should be fixed in source control. But later if castor is updated, and should the updated castor be used to generate source files? The bug fixes already made into an organization's source control need to be checked to see if the latest version of castor fixes those bugs already? If later, the dtd or schema is updated(some elements are changed, new elements are added), castor needs to be run to generate source files from the updated dtd or schema. Then the content of newly-generated files can be different from the existing files. Then the bug fixes to the existing files need to be check to see if it's needed for the new files. I am wondering if this is how it's done in real world. I am concerned that the cost I can save up front can be much less than the maintenence cost. Can I be better off writing my own classes to handle the java xml binding?(no, I don't mean writing a code generator. I mean using a parser to read XML and make some objects out of it, just like the generated classes). I really need a quick response...... Thanks. Kai _________________________________________________________________ �������������ĵ����ʼ�ϵͳ�� MSN Hotmail�� http://www.hotmail.com ----------------------------------------------------------- If you wish to unsubscribe from this mailing, send mail to [EMAIL PROTECTED] with a subject of: unsubscribe castor-dev ----------------------------------------------------------- If you wish to unsubscribe from this mailing, send mail to [EMAIL PROTECTED] with a subject of: unsubscribe castor-dev
