At 11:52 01/07/19 +0200, Patrik F$BgM(Btstr$B�N(B wrote: >In the IDNA proposal we have specified that an application which handle >IDNA (like you seem to say that the browser is) only support IDNA if also >the content of the clipboard is converted to ACE before the domainname is >passed to other applications (which doesn't handle IDNA/ACE solution).
The clipboard doesn't have any protocol to distinguish whether the target will be IDNA-enabled or not. [Most clipboards now have mechanisms to distinguish the encoding, but ACE won't work with that. It's also possible to create new clipboard categories, but again old applications won't pick that up.] Of course, IDNA could require that the clipboard is always in ACE, but then imagine the following: User A sees a cool website, and copies the address (mostly a domain name) into a mail to user B. The paste operation will suddenly show some random garbage rather than the actual address. The user will retry the cut/paste a few more times, with the same result. It will then decide that the browser or the mailer, or both, are buggy, and will type in the address by hand. Also, assume that the clipboard is tagged and that the receiving mailer actually gets the ace and understands it. Should that then go into the mail body as characters, as the user expects? Or would idna prescribe that it goes there as ace? Do we want idna to require behavior that users will clearly see as buggy and not accept, and that therefore reasonable implementers won't implement? And if we take the napkin as another kind of clipboard, is there a way to get the user to somehow change the things she sees to ACE? Is this what we are working on this effort for? In my eyes, clearly not. By the way, what if user B receives the web address in characters (rather than ace)? Well, they'll put it into their browser, and it will either work or not. If it doesn't work, they'll blame the browser. They may find out that they have to upgrade if they want this functionality. To most, it should not be too difficult to understand that to make browsers work with non-ASCII addresses where they have up to now only worked with ascii addresses needs some change somewhere. It's definitely much easier to understand than the clipbord bug, and it's something that users will have to learn anyway, *whatever approach we take*. So my proposal would be: Don't try to be too intelligent with the clipboard. Regards, Martin.
