solves the
issue. It just might take longer for the problem to surface.
Though I'm still not convinced the locking problem isn't a bug. IMHO a
database should never allow itself to be left in a situation from which
it cannot recover after a query.
Nico Coesel
> -Original
gt; like 'COMMIT' can correctly make the changes you want but return an
> error message anyway.
>
Yes, I check every return value. I also see errors from process B saying
the database is locked. In /proc/locks I can see process A still locked
the database. I a
e the locks.
I suspect there may be a path in the code which skips releasing the
locks.
Nico Coesel
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
3 matches
Mail list logo
