I’ve just started using a DataSet at work
and source control is a nightmare i.e. using a sync tool/source comparer to sync the source code to the server. As there are about 4 developers that would be modifying the dataset, the syncing could get quite tricky. Has anyone got any ideas or opinions on this?
We currently use data access classes etc. which makes life easier as changes in code can be easily seen and synced. As with the dataset, its code is auto generated. and if more than one person is modifying it, it’s quite hard to merge the changes…any ideas?
I work with an ORM as well, and our approach is to designate a single person as being responsible for doing all code generation. When a developer needs to modify the schema, they have to contact the “ORM Master”, who applies the schema changes to a clean, “canonical” copy of the database, then re-runs the generation tool, and checks in the new code. The person who requested the change then gets latest and is up and running, and the build doesn’t get broken.
While this adds something of a bottleneck, in practice it has not been a problem, and it avoids issues where two or more developers make different schema changes that add conflicts, etc., causing a knot of problems to sort out.
I hope this addresses the question you are asking, this is how I interpret your question, in any case.