Passing the sqlite3 command from the main process to a child process: is it possible?

I'm experimenting with Tcl to see if a child process can execute without blocking the main process.

I want to pass the dbws command and the channel chan to the child process and have it retrieve the data and write it to the channel without blocking the main process. I realize that there is a -command option that can be used to do this in the background, but I'd like to know if this is possible for other reasons.

Would you please tell me if it is possible to pass a sqlite3 command to a child process rather than having to open a new connection to the same database? I have not been able to get the child process to recognize the sqlite command.

Thank you.

package require sqlite3
if { [catch {sqlite3 dbws -create $db_name false -readonly false} result] } {
  chan puts stdout "result: $result"
  exit
}

puts stdout "Return value of exectest.tcl is:\
 [exec [info nameofexecutable] exectest.tcl dbws $chan]"

The answer is a flat No.

The SQLite library requires the real filename of the database in order to access things like the rollback journal; it doesn't just pass that name to he open() syscall uninterpreted. It is also good practice for connections to not be shared even between threads; some languages do that, but it isn't a good plan. (In particular, starting a transaction from two threads at once in a single connection is Very Bad.)

The Tcl interface to SQLite keeps each connection strictly interpreter-bound (a fact true of all commands), which implies being bound to a single thread of a single process. It does not expose any underlying file descriptors at all, but even if it did, it would be an extremely bad idea to use them. Let each interpreter that needs a database connection open it's own... or delegate the work to a single thread (write heavy workloads are best doing this, and possibly tuning the way transactions work depending on the application use case).