-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 20/01/16 20:11, Roger Binns wrote: > SQLite 3.10 changed the xCurrentTime (note not -64 version) method > for From the documentation it is hard to tell who is at "fault" > here. xCurrentTime is documented as optional now, but was it when I > first wrote the code? It certainly has been present in the "unix" > vfs for many years.
I used the wayback machine to check, and certainly SQLite has now changed in an incompatible way, and will cause null pointer access if you replaced an earlier DLL with the current version (ie no code changes etc). And yes it happened to me. The concept of "shims" - ie one vfs calling another is documented, and hence can be reasonably considered an endorsement of the practise: https://web.archive.org/web/20110626043941/http://www.sqlite.org/vfs.html The xCurrentTime method is not marked as optional: https://web.archive.org/web/20100612193126/http://sqlite.org/c3ref/vfs.html I don't think it is unreasonable that the default sqlite vfs on a platform has to have the xCurrentTime method. Roger -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAlaiggEACgkQmOOfHg372QRUZgCeIBfFpjM4JWOwtTH2Pr9homwg bSYAoLsNzswKudUEtiMhc99V0y2PmBKl =A/pM -----END PGP SIGNATURE-----
