John Stanton <[EMAIL PROTECTED]> wrote: > Have you considered an option where the user can supply the seed so that > an application can implement a complete sequence from the pseudo random > number generator? >
When testing (-DSQLITE_TEST=1) the seed is always 256 bytes of 0x00. That way the same random sequence is generated on multiple runs of the same test. If users need to set their own seeds in a standard build, they can overload the sqlite3OsRandomSeed() interface using the redefinable I/O mechanism. Let's avoid adding yet another API where it is not absolutely necessary. -- D. Richard Hipp <[EMAIL PROTECTED]> ----------------------------------------------------------------------------- To unsubscribe, send email to [EMAIL PROTECTED] -----------------------------------------------------------------------------
