687581 - Can no longer set a cookie exception for local files, and migrated exceptions do not work

Status: RESOLVED INCOMPLETE

(Camino Graveyard :: Preferences, defect)

Description

[Smokey Ardisson (offline for a while; not following bugs - do not email)]

It's unclear whether this is a Camino bug, a Gecko bug, or possibly both.

Back in 1.6.x, when a local file wanted to set a cookie, if you added an exception (Allow/Deny/Session), we(?) stored an exception for "scheme:file".  When you visited another local page that wanted to set a cookie, you weren't re-prompted, per the stored exception.

Now, however (beginning with 2.x), we don't store that exception when you try to set it.  Worse, if you had that exception set in 1.6.x, it gets migrated successfully, but it's not honored.

I know we did a bunch of validation work on exceptions at one point, so it's possible that we broke all or part of this.  It's also possible that the change to mozStorage for cookies somehow broke part of this (although given that cookies are stored in cookies.sqlite and exceptions are stored in permissions.sqlite, that's unlikely).

For users who use lots of local files that set cookies and have "Ask" set (admittedly, not a huge set, but I've noticed it, and one user on the forum is rather upset about it right now), it's annoying.

Comment 1

[Smokey Ardisson (offline for a while; not following bugs - do not email)]

Bleah, I should have spent more time searching.

This was, in fact, intentional (code-wise, though not necessarily behavior-wise): see bug 400097 comment 9, but also bug 204285 (they reused a really old bug to handle the re-breakage from bug 400097).

So, it looks like we're out-of-luck here, except if we wanted to do something bug 204285 comment 15 option 3, which I'm not sure is even possible (consumers = apps, or consumers = code that uses permissions?) :-(

I'll leave this on file for now for tracking/dupe purposes (since it is a valid bug) and also for any "potential" investigation of whether Camino code is also a factor, but that's pretty much moot.

Comment 2

[Smokey Ardisson (offline for a while; not following bugs - do not email)]

(Bleh)^2, bug 204285 is a patch to browser.js, so it doesn't fix the permissions manager at all :/

Comment 3

[Smokey Ardisson (offline for a while; not following bugs - do not email)]

(In reply to comment #0)
> I know we did a bunch of validation work on exceptions at one point, so 

Bug 480892 was what I was thinking of there.