-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 D. Richard Hipp wrote: > If you have or know of a virtual table implementation (other than the ones > that are included with SQLite - such as FTS1-3) that will be broken by an > API change, please let me know. And please also advice me how much of > a hardship a change would be for you.
APSW lets you write virtual tables in Python. The BitPim project is using the functionality, but I have no idea how many others are. Since you marked the API as experimental, I have no problem with even severe changes. (The existing version of SQLite works just fine so noone is forced to upgrade). However please ensure that old code will fail to compile (ie error) rather than have compiler warnings so that it isn't possible for old api calls to be erroneously used with the new api. It would be nice to get some bugs fixed such as 2127. 2128 may even be worth a look - maybe virtual tables could use any SQLite data type as the rowid and not just integers? Roger -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) iD8DBQFHKiVXmOOfHg372QQRAkyWAKDOPFHsNVVc2QGS6a8RIKqmP6mbZwCfdKW8 aIlTPjkIHm8JCQGdKYVqzQk= =NNR/ -----END PGP SIGNATURE----- ----------------------------------------------------------------------------- To unsubscribe, send email to [EMAIL PROTECTED] -----------------------------------------------------------------------------