Same safety caveat applies as in bug 404247. Unfortunately, Privacy can't lose all the core linkage until cookies are similarly wrapped, but this is a step in the right direction.

The patch needs a new nib, since I changed the column identifiers.

Comment on attachment 289228 [details] [diff] [review] fix To mento, per IRC. Just ignore the now-obsolete project stuff.

Comment on attachment 289228 [details] [diff] [review] fix This will bit-rot bug 354975 a little, but that's OK. The nib is the part that's a real pain, can you just confirm that the only thing you're changing is the column identifier? If it turns out that this is bit-rotten relative to the current trunk code and this needs to be updated, maybe we should land 354975 first. If you can apply this, go ahead, and I'll worry about 354975. Too bad we still need the Gecko dependencies for nsICookieManager and nsICookie. + [permManager setPolicy:inPermission //XXX What's this? Looks good.

Yep, column IDs are the only change, so it'll be easy to resolve the nib conflict in either order. As for nsICookie*, that's definitely part of my master plan... I just didn't expect this initial permission stage to take months while the project got sorted out ;) The comment was cruft from memos to self during rewriting, and will be removed at checkin.

Comment on attachment 289228 [details] [diff] [review] fix rs=pink

This is what I'll actually check in on the branch; gcc-3.3 gave me some grief. Trunk version is this minus the asm bits.

Landed on trunk and MOZILLA_1_8_BRANCH.