On Sun, 3 Aug 2014 18:04:35 -0700 Marc MERLIN <marc_...@merlins.org> said:

> On Mon, Aug 04, 2014 at 09:02:16AM +0900, Carsten Haitzler wrote:
> > > Since I upgraded to the newer e18, Multi doesn't work anywhere except in
> > > efl apps it seems (see below)
> > 
> > multi key compose? no - it works. i use it regularly. as i said - i spent a
> > lot
> 
> Yes. It now works only in E and terminology for me.

it's directly supported in efl core - i believe for gtk, etc. you need input
methods for this (scim/ibus/iimf etc.). i use those too, so it works there for
me. in efl there is in fact a compiled-in table of all of the compose sequences
for multi key. terminology uses that api to figure out your compose seq, and
edje entries (which elm entry uses) do too. thus it worksin efl.

> > of time looking at this. the modifier masks for alt and altgr clash. they
> > are indistinguishable from a client point of view. it's a misconfiguration
> > of the modifier masks.
> 
> So wait, did you say earlier the problem was with the X server (code?)
> or my configuration of xmodmap?

xmodmap creates a conflict of modifier maps with your setup. phab ticket has an
explanation - as does the git commit log.

> I have this (default from xorg), should I change it?
> legolas:~$ xmodmap -pm
> xmodmap:  up to 4 keys per modifier, (keycodes in parentheses):
> 
> shift       Shift_L (0x32),  Shift_R (0x3e)
> lock        Caps_Lock (0x42)
> control     Control_L (0x25),  Multi_key (0x69)
> mod1        Alt_L (0x40),  Mode_switch (0x6c),  Meta_L (0xcd)
> mod2        Num_Lock (0x4d)
> mod3      
> mod4        Super_L (0x85),  Super_R (0x86),  Super_L (0xce),  Hyper_L (0xcf)
> mod5        ISO_Level3_Shift (0x5c),  BadKey (0xcb)
> 
> Actually I'll attach my xmodmap dump so that you can tell me if I have
> something set wrong or if there is something wrong with my x server.
> More simply, if you can tell me what I should do to assign modeswitch to
> and multi to right alt and ctrl, I'll be happy to oblige :)

as above. i don't know how to fix your modifier map other than somehow
assigning mode_switch to something other than mod1 - maybe mod3. how is a
matter of tinkering.

> > > I see this even when when I have 108 set to ALT_R and not Multi.
> > 
> > err what do you mean by "multi"? there is an actual multi key. this is how
> > you do compose sequences. ie:
> > 
> > multi, e, ` == è
> > multi, e, ' == é
> 
> I have multi bound to right ctrl, it's worked like this for about 20
> years. It still works with terminology and e, but nothing else now.
> 
> I don't need multi for é since I can use mode_switch to hit right alt
> and e, but for ë I do use 
> right ctrl
> "
> e
> 
> It still works in terminology, but nowhere else now.
> OMG, I ran a new xfce4-terminal, and now mode_switch and multi work in
> it, as well as all my existing xfce4-terminal windows. WTH?
> As the same time, mode_switch now broke in terminology, but multi still
> works in it.

what are you doing?

> > you get the above as per all my comments in the phab bug. i am repeating
> > myself now endlessly. go read it. i explained it in detail. they way you
> > have your keyboard working alt and altgr as modifiers conflict. they
> > produce the SAME modifier mask. exactly the same. modifier state is a
> > bitmask - with a bit on for the modifier. e dumps altgr if it conflicts
> > with alt as it is not a separate modifier at all then, and that is what
> > causes the problem with bindings as e can now not tell if its alt or altgr
> > pressed, so it assumes both as they both match - otherwise it has to
> > randomly guess which and that's worse.
>  
> I understand that part, but I think I misunderstood you on where the bug
> is. I thought you meant the bug was in X, but now I think you're saying
> I have my modifiers set wrong in xmodmap, and that I can change of
> course.

you have altgr create a modifer mask of 0x8 (bit # 4 set). alt ALSO produces
this bitmask as well. (i don't remember the exact bit bit they both had the
SAME mask). since we detect modifiers by the bitmask - it's now ambiguous.

try it. hit alt+x then altgr+x - in xev. look at state value. that is the
modifier/lock states. notice it has the same value with either key. note that
with bindings you dont see the modifier key press. you only see the "x" or "f1"
or whatever along with the current state, so we can't go tracking the modifier
keys themselves.

> Note that I don't knowingly change my modifiers (xmodmap -pm that is, I
> did change the key mappings, aka xmodmap -pke) with xmodmap, I think
> that's what I get from X, but I'm happy to adjust them.

you implicitly do. unfortunately.

> > xmodmap -e "keycode 108 = Mode_switch Alt_R Meta_R Alt_R Meta_R"
> 
> I now simply have
> keycode 105 = Multi_key
> keycode 108 = Mode_switch
> Is that broken?
> 
> > doesn't even do anything for me in 2 apps as non-efl toolkit apps. so i
> > don't even know what it's meant to do/make work... but it does nothing for
> > other toolkits, and nothing for terminology or efl too. all it results in
> > is a broken modifier mask mappping that i know have added code to work
> > around.
>  
> There is clearly something subtle going on since it's worked for me for
> years with everything all the way back to Xfree286, but I'm not attached
> to keeping in that way, I'm happy to change.
> Is it broken to assing Multi to 105 and Mode_switch to 108?

it's not broken because no one allowed altgr as a modifier. like alt, ctrl,
shift. we do that. this creates a conflict.

> On Sun, Aug 03, 2014 at 08:40:59PM -0400, John Holland wrote:
> > éÉ - from Mode_switch/xmodmap
> > 
> > èÈëËêÊ  -- from setting Caps Lock to be Multi in the E settings and
> > typing sequences like capslock,e,backtick   as per Rasterman's message.
> > 
> > Both of these methods are working for me in terminology and claws mail.
> > 
> > The modmap things I found online seem to show that having more than 4
> > values for a key is not well supported, but your (Marc's) message had
> > that tantalizing set of eight values for the e key. I don't see how to
> 
> I don't set 8 values, but xmodmap -pke returns 8 values.
> 
> > this is what I get when I start terminology (my computers are sometimes
> > named after jazz musicians)
> > 
> > jholland@herbie:~$ terminology
> > ERR<3699>:ecore_x lib/ecore_x/xlib/ecore_x.c:301
> > _ecore_x_modifiers_get() ALTGR conflicts with other modifiers. IGNORE
> > ALTGR
> 
> Ok, so you're also getting this conflict without my Xmodmap, but at
> least things seem to be working for you. Thanks for sharing.
>
> Marc
> -- 
> "A mouse is a device used to point at the xterm you want to type in" - A.S.R.
> Microsoft is to operating systems ....
>                                       .... what McDonalds is to gourmet
> cooking Home page: http://marc.merlins.org/                         | PGP
> 1024R/763BE901


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    ras...@rasterman.com


------------------------------------------------------------------------------
Infragistics Professional
Build stunning WinForms apps today!
Reboot your WinForms applications with our WinForms controls. 
Build a bridge from your legacy apps to the future.
http://pubads.g.doubleclick.net/gampad/clk?id=153845071&iu=/4140/ostg.clktrk
_______________________________________________
enlightenment-users mailing list
enlightenment-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/enlightenment-users

Reply via email to