On Nov 11, 2009, at 1:24 AM, Scott Hess wrote: > Someone was asking me a question about what happens if the wrong > journal file gets applied to a database. My answer was that terrible > things happen. > > I found myself wondering if the "File change counter" value is stored > in the journal (*).
It is not. http://www.sqlite.org/fileformat.html#journal_header_format Dan. > It would seem like a cheap(ish) way to give some > reasonable assertions about whether the journal belongs to the > database file it's being applied to. But if it's in there, I'm not > having any luck finding it. Has anyone already figured out that this > is true? > > Thanks, > scott > > (*) It would be more complicated than storing the original value. > You'd probably have to store both the original and the target. And > that assumes that at rollback, the database file's header was either > completely overwritten or not written at all, which may not match > reality very well. > _______________________________________________ > sqlite-users mailing list > sqlite-users@sqlite.org > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
