On 13 Aug 2013, at 1:18pm, Jan Slodicka <j...@resco.net> wrote: > The same NOCASE implementation is used for all database operations. > > I am aware of the consequences of an incorrect collation implementation. > However, I do not believe that they could explain index reference to records > that were deleted. Do you have any explanation for this problem?
Usually the corruption sneaks into the database and remains undiscovered for a while until one particular command fails because of it. Does your code check the results returned by all operations on the database, even when they 'must' be SQLITE_OK ? Simon. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users