[Webkit-unassigned] [Bug 13471] REGRESSION (r21045): Secure keyboard entry mode remains in effect after leaving a password field by submitting

[bugzilla-daemon at webkit.org]

http://bugs.webkit.org/show_bug.cgi?id=13471





------- Comment #5 from darin at apple.com  2007-04-24 09:47 PDT -------
(In reply to comment #4)
> Why not say "you made a typo"? ;-)

Sorry, I was just in a silly mood.

> > What's the key reason this needs to be called? Is it really the change in view
> > that's the issue? Doesn't the input element lose focus?
> 
> In many senses, it never does (so I think that even prior to r21045, if you
> happened to be in a secure text field when the new page was loaded, you'd
> remain in Roman-only mode). The Frame doesn't keep track of focus, it just
> serves requests from nodes to switch secure mode on or off. Presumably, those
> nodes are in the Frame's Document. So now I'm thinking it may be more
> appropriate to patch setDocument() instead of setView().

Sounds good; easier to explain at least. I think either is fine in practice.

> It would still just
> reset to false unconditionally, since it can't (and doesn't need to) ask the
> new document's focused node if it requires secure keyboard entry. However, that
> could be changed in the future (if password fields can go into the page cache).

Makes sense. This is subtle enough that I think we need a comment or two. I
really like the idea of putting a comment in FrameLoader::canCachePage -- I
don't think it needs to be a FIXME, just a warning that if you change the rule
then this other code will need to change too.


-- 
Configure bugmail: http://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.