On 5/20/15, Simon Slavin <slavins at bigfraud.org> wrote: > It has been an extremely strange week for database corruption and extremely > complicated bugs. I'm going to have to move "try PRAGMA integrity_check" up > a few notches in future. Although, as Richard comments, one of the reports > was of corruption that integrity_check didn't spot. >
Two corruption bugs in two days. Ugh. On the other hand, neither bug resulted in loss of data. Both were formatting issues only. And in both cases, the SQL needed to tickle the bug is rather obscure. Note also that the second problem (the TEXT in an INT column) is *not* currently detected by PRAGMA integrity_check. That is something we might address in the future... -- D. Richard Hipp drh at sqlite.org