--- Ivan Pascal <[EMAIL PROTECTED]> schrieb: 
> > I noticed when working with the right Alt key as a
> mode-shift
> > key that it matters whether Shift is pressed after or
> before
> > the right Alt-key. In order to get the correct behaviour,
> I
> > have to press the Alt key first, then the Shift key. The
> > other way round leads to the Alt key getting ignored.
> This is
> > different from how the keys behave under Windows. Is this
> > intentional or is it a bug?
> 
> I's a silly decision in a keboard map.
> In many maps this AltGr (Right Alt) key has two keysyms:
> <mode shift> and
> Multi_key.
> 
> (Multi_key is a special non-printable symbol used as a
> prefix in some composing
> sequences.)
> 
> When one presses this key along it produces <mode shift>
> keysym but being
> pressed with Shift (when Shift is already pressed) the key
> omits Multi_key
> keysym.  There is nothing wrong with it if all keys in a
> map have no more than
> three symbols (without modifiers, with Shift, with <mode
> shift>).  But if a key
> has fourth symbol (that should be accessable with
> Shift+<mode shift>) a getting
> such symbol is unhandly becouse it depends on the order of
> the Shift and
> <mode shift> keys pressind.
> 
> I don't know who put two different keysyms on AltGr key
> first.  But now such
> configuration is copied in many keyboard maps.
> 
> The solution is obvious: remove Multi_key keysym from AltGr
> key and tie it
> to some other key.  And there are a few keyboard map
> authors an maintainers
> who have already done it.
> But there is not any standard key for Multi_key.  Somebody
> puts it on RightWin
> key some others on Menu.  And where would be a good place
> for Multi_key in
> a keyboard without 'Microsoft keys'?
> 
> You see, the problem can be fixed easy but as for me I
> don't know where people
> are used to have this keysym.

Thanks for your clear explanation. I was in the process of
putting together a keymap that implements the
US-international keyboard mapping available in Microsoft
Windows. I copied the setting for the right Alt key from
other keymaps, just as you said, not realizing what it did.

> > There's a minor niggle left, however: Under Windows the
> > behaviour of the dead keys is somewhat different. When a
> dead
> > key is followed by a kepress with which it doesn't
> combine,
> > the character of the dead key is generated followed by
> the
> > character of the key pressed thereafter. For example if I
> > press the apostrophe ' followed by the letter p the
> result is
> > 'p. Under X the apostrophe is lost and just the p is
> printed.
> > If I press the apostrophe twice under Windows then two
> > apostrophes are emitted, whereas under X a single
> apostrophe
> > is emitted. Is there a way to fiddle with this behaviour
> > under X to make it work in the same fashion as under
> Windows?
> 
> It's a problem.  The thing is that X Window doesn't have a
> separate module
> for dead keys processing.  Instead of it X has a 'Compose
> mechanism' which
> allow to compose a symbol from any sequence of keys.  Since
> such sequence can
> be relatively long this Compose module doesn't remember all
> keys already
> pressed (even it is a one key only) but keeps some
> 'composing state'.  When
> the module gets a key event that doesn't match this state
> (can't be a
> continuation of a 'compose sequence') all what it can do is
> to give this
> symbol to an application (and reset the state).  Since it
> doesn't remember
> a previous key it can't give such symbol together with the
> last one.
> 
> If there are few combinations you really want (two
> apostrophes and something
> else) you can just add them into compose rules. (All
> compose rules are written
> in a special "Compose" file in a human readable form and
> can be changed easy
> using any text editor.  Also note that the result of
> composing isn't limited
> to one symbol only, it can be a string of symbols.)
> 
> But if you want such beahaviour for any combination of keys
> it requires
> a rewriting (redesign) of the composing module.  And I
> think there are many
> people who are used to the current behaviour and will
> consider any other one
> as unexpected and unhandly.

I won't go that far, that's way too much work for what it's
worth. My US-international keymap will have to do without it.

Thanks again for your help.

Cheers
Stefan


__________________________________________________________________

Gesendet von Yahoo! Mail - http://mail.yahoo.de
Logos und Klingelt�ne f�rs Handy bei http://sms.yahoo.de
_______________________________________________
XFree86 mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/xfree86

Reply via email to