> So, OK to remove? One more question: The feature that selects into a file on client side, where is that located?
If it's in the drizzle/mysql client app, then there is a problem that for a developer who uses the library, this feature is essentially removed completely. If it's in the library, then the followup question is, what about the PHP, JDBC and other libraries? (Including MySQL libraries in this survey is ok.) Even if the C library supports something, others might not always. The point of this question is that moving the server side code is ok, if this is still possible to do in all these clients side use cases: drizzle/mysql client app, C client library, PHP client library, JDBC client library, other client libraries... If it becomes completely impossible to dump into a CSV file for some of these use cases, the I think you could still remove the server side code, but create a blue print about adding the feature back to those libraries where it is missing. (It's not a critical thing, say it is missing in the JDBC driver, then a Java developer can of course create a CSV file himself. But it's still convenient to have a method that does it for you, and does it in a standard output format.) Clint's point about security is imho the best argument for removing it. It's true the MySQL feature is pretty well limited, but the existence of the code is still a potential source for bugs, which lead to security breaches. (For instance, in Drizzle we don't have similar grants as in mysql, so anyone can do this now? I don't know, but it scares me to ask...) henrik -- [email protected] +358-40-8211286 skype: henrik.ingo irc: hingo www.openlife.cc My LinkedIn profile: http://www.linkedin.com/profile/view?id=9522559 _______________________________________________ Mailing list: https://launchpad.net/~drizzle-discuss Post to : [email protected] Unsubscribe : https://launchpad.net/~drizzle-discuss More help : https://help.launchpad.net/ListHelp
