On 24 Jan 2012, at 6:43am, David Henry wrote: > I am working without an operating system so there are no other agents trying > to steal data.
You're writing an operating system ? That can be lots of fun. Good luck. > Bearing that in mind, is it still necessary to actually write > zero data to the sectors allocated? Is SQLite expecting it? I haven't looked though the code with that in mind, but as far as I know, SQLite does not make any assumption about what will be in newly-assigned sectors. If it wants zeros there it'll write them itself. It is ftruncate itself which chooses to write zeros to any newly-assigned pages of disk space. This is part of the specification of ftruncate, and documented here: <http://pubs.opengroup.org/onlinepubs/007908799/xsh/ftruncate.html> If your version of ftruncate doesn't write the zeros, you haven't implemented ftruncate properly. But if you're not trying to reproduce UNIX, I guess it doesn't matter. Simon. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
