Re: [Fwd: Re: New string in gnome-control-center]

[Danilo Segan <danilo gnome org>]

Jody Goldberg <jody@gnome.org> writes:

> On Fri, Feb 27, 2004 at 03:03:57PM +0100, Danilo Segan wrote:
>> 
>> What about the those non-technical, XKB equipped, long-standing X
>> users?  ...
>
> Why would they see the message at all ?  They wouldn't have
> .xmodmap* files lying around.  

Oh, I missed to repeat "using xmodmap" there, since those were the
users you seemed to have been targeting with this message, yet it is
my opinion that they won't be able to fix the problem even with this
message, so it's unnecessary.  Technical users may peek into
~/.xsession-errors (I believe that was the reason you mentioned users
"not necessarilly technical", though I may be wrong).

> There is no way to use the old modmap files at this time.  Our
> current feeling is that the message should be a simple notification
> so that the user is at least informed that the file will be ignored.

That's the core of my point: they'll soon realize that their settings
don't work.  They just need to press a couple of keys on the
keyboard, and it doesn't take too much intelligence to realize that
your previous keyboards settings suddenly don't work.  This
notification is being nice to user, but I still don't think it's
essential in order to break the string freeze.

I don't think this will stop bug reports either -- people will come
with idea that it should still work, and that they're seeing a
"stupid" message.

> Remember, searching for these file names is _not_ conclusive, it is
> just a curtesy.  Someone that put the modmap commands directly into
> their .xsession would not get notified.

Yes, I realized that from the patch, and that's one more reason to
*not* show this message, since it's only a heuristic which might not
work in many cases (that's why I talked repeatedly about ~/.xmodmap
file, not only "xmodmap").  If we want to be nice, lets be nice to
all our users, not just some lucky ones.

> There is no fix.  Disabling xkb in the server would do it, but is
> pointless.  The user should be able to reproduce their cherished
> modifications using the extensions to the keyboard capplet.

I know, perhaps that's what should be suggested in the dialog (if
it's at all showed), instead of the currently proposed message: a user
finding out about the way to fix it will appreciate it.

> In summary, we suck and should not be erasing a user's
> modifications.  However, since there is no current way to avoid it
> we need to at least warn them.

That's where we disagree.  I'll point out again that I don't see much
good in issuing a warning, since the user is going to find that out
very soon on their own (if not already, imagine a user having switched
"I" and "R", and those are shortcuts on the "_Ignore" and "_Rename";
ok, very unlikely, but this user might do what he didn't intend by
pressing Alt+R instead of Alt+I).

Of course, since you're doing the work here, I won't mind any
decision you come up with, but these are the points that should be
taken into account, IMHO.

Thanks,
Danilo