On Fri, Nov 13, 2009 at 01:15:40PM +0100, Alexander Sack wrote: > On Thu, Nov 12, 2009 at 03:13:19PM -0800, Josh Triplett wrote: > > On Thu, Nov 12, 2009 at 11:59:39PM +0100, Mike Hommey wrote: > > > On Thu, Nov 12, 2009 at 02:49:44PM -0800, Josh Triplett wrote: > > > > On Thu, Nov 12, 2009 at 11:47:01PM +0100, Mike Hommey wrote: > > > > > On Thu, Nov 12, 2009 at 02:42:12PM -0800, Josh Triplett wrote: > > > > > > > Yes, but there is a remove all cookies in the cookies list. If > > > > > > > you do > > > > > > > that in a new profile with new cookies, are they properly > > > > > > > "securely" > > > > > > > deleted ? > > > > > > > > > > > > I just tested that, and no, that doesn't work either. Starting > > > > > > from no > > > > > > data in cookies.sqlite, I visited google.com, closed Iceweasel, > > > > > > confirmed that a couple of cookies appear in cookies.sqlite, opened > > > > > > Iceweasel, used "Remove all cookies", closed Iceweasel, and > > > > > > confirmed > > > > > > that the cookies still appear in cookies.sqlite even though > > > > > > moz_cookies > > > > > > has no rows. > > > > > > > > > > That's really weird, because there is nothing I can see at first hand > > > > > in > > > > > the mozilla code that differs between the various .sqlite files. So it > > > > > looks like this could be a sqlite bug. Have you tried to remove the > > > > > lines with sqlite3 itself ? > > > > > > > > Yes, "sqlite3 cookies.sqlite vacuum" does wipe the data. > > > > > > Except that vacuum is not the one, you should try a delete statement. > > > (vacuum is not used in mozilla, except in expiry of form history, when > > > it is big) > > > > Oh, I see. > > > > No, "delete from moz_cookies" in sqlite3 doesn't securely wipe the data > > either, though a subsequent "vacuum" does. > > hmmm ... could you check that this works properly with upstream firefox > builds?
(One installation of ia32-libs-gtk later, which still doesn't quite work right due to various missing GTK+ helper libraries...) No, upstream firefox doesn't get this right either. I'll file an upstream bug and mark this as forwarded. - Josh Triplett -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
