Re: [evolution-patches] Seeking review for bug #47751 & #47752

[Ettore Perazzoli <ettore ximian com>]

On Wed, 2003-08-20 at 12:19, Anna Marie Dirks wrote:
> (Purely from a usability perspective, not thinking about string freeze
> atm) There is a delicate balance to be struck between 1) warning the
> user too much, hence annoying her and 2) not warning the user enough,
> hence letting her lose her data accidentally. My recommendation is
> that in cases like the druids -- the import druid, the ldap server
> druid, the pilot set-up druid and the mail account druid-- where the
> user has to input several pages of (potentially confusing and hard to
> remember) data, it is good practice for us to show a confirmation
> alert when the cancel button is clicked, or the esc button is
> pressed.  
> 
> In other cases, like the filter editor, and the vfolder editor, the
> user doesn't really risk losing very much data if she accidentally
> closes the dialog. In these cases, it is probably more trouble than it
> is worth to show the user a confirmation alert. 

I agree.

> Since we obviously don't live in such a world, maybe any new alerts
> that you guys add, Maxx, should only go into HEAD for now? 

Yes.

So, aside from the UI guidelines, the sum is (just to clarify):

      * No patches adding confirmation dialogs should go into HEAD.  (Or
        any dialogs, for that matter.)

      * This confirmation dialog doesn't go in at all.

Also, when you submit a patch, please make it clear where the patch is
supposed to go (HEAD or stable).  I have had a bit of trouble figuring
it out myself lately.

For stable patches, we must limit ourselves to the bare minimum: data
loss, crashes, security issues, completely broken functionality and so
on.  Less critical fixes should only go to HEAD.

Thanks,
-- Ettore