On Fri, Mar 7, 2014 at 5:27 PM, Eleytherios Stamatogiannakis <est...@gmail.com> wrote: > IMHO, SQLite is targeted towards being a relational "core" with very wide > extensibility. Adding specific scripting engines to it is would be > detrimental to its main purpose (being a very good relational "core").
You are misrepresentation the discussions in this thread IMHO. We never discussed adding a scripting engine to SQLite, but making it easy to supercharge SQLite with the ability to write extension function via scripting (in the broad sense) instead of via compiled C extensions. And we proposed a syntax to declare those scripted custom function in a way that's fully compatible and consistent with the way virtual tables are declared. And once again, all of the above is already possible, via compiled extension modules, as Clemens pointed out, and you also pointed out by mentioning your own Python-implemented custom functions. But you had to write custom C code to be able to have SQLite know about and use your Python-implemented custom functions, and you likely embed SQLite inside your Python environment, while I propose the reverse, and a standard syntax to register those functions, which is extension-module and thus scripting-engine independent. In any case, Dr. Hipp does not participate in this thread, so it will likely stay wishing thinking :). I still think my proposal is elegant, SQLite-esque, useful, and importantly doable; and lightweight in code size to boot. But of course documenting and testing this, to SQLite's high standards, that's definitely work. --DD _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
