Marco Maggi <marco.maggi-i...@poste.it> wrote: > I am a True Beginner with SQLite and I am writing a > binding to it for a programming language. While wrapping > the functions accepting "sqlite3_value" arguments, I am not > sure if I understand from where "sqlite3_bind_value()" (and > "sqlite3_result_value()") should take their arguments.
Your custom function is passed an array of sqlite_value* pointers as a parameter. If it wants to return one of these parameters without modification, it would use sqlite3_result_value. If it wants to run a parameterized statement and use one of its parameters as a parameter to the statement, it would use sqlite3_bind_value. > Is it actually possible and safe to execute a statement > from the implementation of an app-defined SQL function? Yes. > So SQL function arguments can be bound to statement parameters, > and results from a statement can be used as return values of > app-defined SQL functions? Yes. > Can I think of instances of "sqlite3_value" as shared data > structures whose life is correctly handled by SQLite with > some mechanism like reference counting? sqlite3_value structures passed to your function are only valid until the function returns (sqlite3_result_value of course takes a copy as necessary). > * Is there some simple and "known" example of such mechanism > I can put in my test suite and documentation to show how > it is done? Example of which mechanism? > * If the nested statement execution fails for some reason: > is there some convention about how the app-defined SQL > function should signal the error to its caller? sqlite3_result_error* -- Igor Tandetnik _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
