Hi, Jyri Virkki wrote: > Thanks, most of the updates look good. A few more comments: > > >> This is test programs that comes with the standard distribution, but >> they are not necessary to run the MySQL server and other binaries. >> They are necessary to ensure that our way of building MySQL does not >> break any functionality. >> > > So these are only for build testing (the typical 'make test' cycle)? > If yes, is there really a reason to package them? > Or are they useful for end users? > I think they could be useful for users as well, both for running sql-benchmark test as well for examples of how to use the client api's. On the other hand I think most users are quite familiar with MySQL and are just interested in the server and the admin and utility tools. > >> It is the MySQL server binary. The interface should maybe be the "server >> program command line options" ? >> > > How do I pass those command line options when starting via smf? If > these are needed, they probably need to be smf properties? Or can they > be in config file? > All options for the MySQL server can be in the default config file, also the datadir option, so we don't have any use for passing the command line options to smf, but it might be that users want to start the MySQL server without smf to test different configurations of MySQL f.ex. > >> ok, error codes should be Commited interface while the messages can >> change as you say, so they should be UnCommited. So scripts should rely >> on the errorcodes and not messages. Should both be listed here as >> interfaces? >> > > Note that Uncommitted is also a relatively stable public interface > type. Volatile is closer if the messages can change, but if scripts > should not rely on the exact message text anyway, it is "Not-an-interface". > > For more background see > http://www.opensolaris.org/os/community/arc/policies/interface-taxonomy/ > > ok, will state the messages as Not-an-interface, and keep the error-codes as Commited.
Thanks, Jan S
