6:52 a.m.
The following has been an annoyance for a while, but now I decided to speak
up :-)
I have a habit of making the useless caps lock key to act as control key,
which I do in .Xmodmap. I also do things with Alt_R and such.
When XEmacs starts up, it gives me a mouthful. Why does it care?
Especially the last two warnings look silly in view of the fact that it is
precisely my intention that capslock would generate modcontrol (and that
the way to do this is in the X windows FAQ).
--michael
(1) (key-mapping/warning)
The meanings of the modifier bits Mod1 through Mod5 are determined
by the keysyms used to control those bits. Mod1 does NOT always
mean Meta, although some non-ICCCM-compliant programs assume that.
(2) (key-mapping/warning)
The semantics of the modifier bits ModShift, ModLock, and ModControl
are predefined. It does not make sense to assign ModControl to any
keysym other than Control_L or Control_R, or to assign any modifier
bits to the "control" keysyms other than ModControl. You can't
turn a "control" key into a "meta" key (or vice versa) by simply
assigning the key a different modifier bit. You must also make that
key generate an appropriate keysym (Control_L, Meta_L, etc).
(3) (key-mapping/warning) XEmacs: Caps_Lock (0x6d) generates ModControl,
which is nonsensical.
(4) (key-mapping/warning) XEmacs: Caps_Lock (0x42) generates ModControl,
which is nonsensical.
7:20 a.m.
>>>> "Michael" == Michael Kifer
<kifer(a)cs.sunysb.edu> writes:
Michael> The following has been an annoyance for a while, but now I decided to speak
Michael> up :-)
Michael> I have a habit of making the useless caps lock key to act as control key,
Michael> which I do in .Xmodmap. I also do things with Alt_R and such.
Michael> When XEmacs starts up, it gives me a mouthful. Why does it
Michael> care?
What you are doing is not ICCCM-compliant, I believe.
The ICCCM says:
Convention
Clients should determine the meaning of a modifier bit from the KeySyms
being used to control it.
which seems to mean that xemacs should interpret your xmodmapped
capslock key with its _original_ meaning.
You have to have one of the Control keys generate control.
You want to do something like this: (editing required)
xmodmap \
-e 'clear lock' \
-e 'clear control' \
-e 'keycode 126 = Control_L' \
-e 'add control = Control_L' \
-e 'add control = Control_R' # if you have a Control_R key
But users typically don't figure this out, and what you're doing
typically works even though it's illegal, so maybe we should turn the
warning off.
Has this been FAQized? If not, it should be.
7:38 a.m.
On Thu, Jan 18, 2001 at 01:52:34AM -0500, Michael Kifer wrote:
When XEmacs starts up, it gives me a mouthful. Why does it care?
The semantics of the modifier bits ModShift, ModLock, and ModControl
are predefined. It does not make sense to assign ModControl to any
keysym other than Control_L or Control_R, or to assign any modifier
bits to the "control" keysyms other than ModControl. You can't
turn a "control" key into a "meta" key (or vice versa) by simply
assigning the key a different modifier bit. You must also make that
key generate an appropriate keysym (Control_L, Meta_L, etc).
OG.
8:07 a.m.
>>>> "OG" == Olivier Galibert
<galibert(a)pobox.com> writes:
OG> On Thu, Jan 18, 2001 at 04:20:29PM +0900, Martin Buchholz
OG> wrote:
> Has this been FAQized? If not, it should be.
OG> What should be FAQized, the fact that XEmacs spews you a
OG> warning, it may be useful to actually read it?
Give Jamie the last word, otherwise you'll regret it.
http://www.jwz.org/gruntle/autobogotification.html
--
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091
_________________ _________________ _________________ _________________
What are those straight lines for? "XEmacs rules."
8:38 a.m.
>>>> "MB" == Martin Buchholz <of Thu, 18 Jan
2001 16:20:29 +0900> writes:
>>>> "Michael" == Michael Kifer
<kifer(a)cs.sunysb.edu> writes:
Michael> I have a habit of making the useless caps lock key to act as
Michael> control key, which I do in .Xmodmap. I also do things with Alt_R
Michael> and such. When XEmacs starts up, it gives me a mouthful. Why
Michael> does it care?
MB> What you are doing is not ICCCM-compliant, I believe.
MB> You have to have one of the Control keys generate control.
Why? I have already assigned control to caps lock. I don't have keys to waste.
I actually bind the ctl key (on the PC eyboard) to save-some-buffers !
MB> You want to do something like this: (editing required)
MB> xmodmap \
MB> -e 'clear lock' \
MB> -e 'clear control' \
MB> -e 'keycode 126 = Control_L' \
MB> -e 'add control = Control_L' \
MB> -e 'add control = Control_R' # if you have a Control_R key
I do something similar:
remove Control = Control_L
remove Lock = Caps_Lock
keysym Control_R = Caps_Lock
add Control = Caps_Lock
keycode 37 = F22
It is probably the last line that causes the warnings: I make this key look
like F22 and then bind F22 to save-some-buffers.
MB> and what you're doing typically works even though it's illegal,
I think it should be illegal to sell keyboards that have Control in the
wrong place :-)
MB> so maybe we should turn the warning off.
Please do. Why should you care how the keys are mapped? Chances are that if
I did something with the keys that can seriously break XEmacs, other
editors won't work as well (e.g., vi).
--michael
12:41 p.m.
>>>> "Michael" == Michael Kifer
<kifer(a)cs.sunysb.edu> writes:
>>>> "MB" == Martin Buchholz <of Thu, 18 Jan 2001 16:20:29
+0900> writes:
Michael> When XEmacs starts up, it gives me a mouthful. Why
Michael> does it care?
Because XEmacs does very complex things with keymapping given the
chance.
MB> What you are doing is not ICCCM-compliant, I believe.
MB> You have to have one of the Control keys generate control.
Michael> Why? I have already assigned control to caps lock. I
No, you haven't (I'm deliberately interpreting control -> Control
keysym, as opposed to control -> control modifier bit) . That's not
your fault; X11 key mapping is brain dead. (Well, actually, to be
fair it was designed before our ancestors had evolved forebrains.)
Michael> don't have keys to waste. I actually bind the ctl key
Michael> (on the PC eyboard) to save-some-buffers !
Martin isn't speaking clearly. What he meant is "you have to have a
key whose keysym is Control_[LR] generate the control modifier." If
you have hands like Akeem Olajuwan, you could put the control modifier
bit on KeypadMinus---but you also need to put a Control keysym there.
Michael> I do something similar:
Not really.
Michael> remove Control = Control_L
Michael> remove Lock = Caps_Lock
Michael> keysym Control_R = Caps_Lock
So far, so good. You now have no key that can do lock, and you have a
key in the lower right corner that does control. That key generates
the keysym Caps_Lock. The key in the lower left corner generates the
keysym Control_L, and the key to the left of the "A" generates the
keysym Caps_Lock.
Michael> add Control = Caps_Lock
You have added the control modifier to the key to the left of "A" and
(redundantly) to the key in the lower right corner.
Michael> keycode 37 = F22
I assume this keycode corresponds to the key in the lower left corner?
That key now generates the keysym F22. You have no Control keys
anymore, although you can generate control keystrokes. Summarizing:
Physical Keysym Modifier
------------------ --------- --------
Left of "A" Caps_Lock control
Lower left corner F22
Lower right corner Caps_Lock control
Michael> It is probably [keycode 37 = F22] that causes the warnings
Nope. The last line is completely unexceptionable. Look at the
errors: 37 != 0x6d or 0x42.
MB> what you're doing typically works even though it's illegal,
It depends on whether it matters if you have a Control key or not. It
probably doesn't. Unless you're Ben Wing, and can't hold down Control
and another key simultaneously on very bad days. Most of us aren't
Ben, and can chord, so we typically think "it works."
Howevva ... some implementations of menu accelerators "do something"
(like highlight the menu bar) as soon as you hit the Alt key. This
will break, for everybody, if there is no Alt key. Doesn't matter
whether or not there is an Alt modifier that allows you to produce
keystrokes like Alt-F.
Michael> I think it should be illegal to sell keyboards that have
Michael> Control in the wrong place :-)
Caveat emptor. :-)
MB> so maybe we should turn the warning off.
Michael> Please do.
Please don't. I'll have to revert it: Jamie knows my email, and I
don't look good as he does in "black and crispy." If you have to do
something, distribute xkeycaps(1X).
http://www.jwz.org/xkeycaps/
Michael, it's not your fault that X11's keyboard model is braindead.
You also (probably; 200 keyboards is a _lot_ of keyboards[1]) don't
have to do anything about it personally, thanks to the fact that the
halflife of Jamie's brain is measured in months (or maybe years).
<RANT>
We, on the other hand, are going to undoubtedly be dealing with
braindeath of the sort that caused the (now obsolete and removed,
praise be) Mandrake-bashing entries in PROBLEMS, because Linux (and
probably the BSDs) are spawning efforts to standardize and debase^W
make user-friendly all user interfaces. For years.
It is _not_ true that XEmacs is no more complex than any other app
(for God's sake Martin, you're the Bullshit/Defeat guru for this
project! you should know better). And it's supposed to get more
complex (have you forgotten that Ben plans to add menu accelerators
for keyboards that have more than 1 extra modifier key?)
</RANT>
Footnotes:
[1] And none of them are JIS, and only about 2 know anything about
Japanese. But that's OK, XFree86's (and Linux's) Japanese keyboard
data are all broken anyway. Another day, another <RANT>.
--
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091
_________________ _________________ _________________ _________________
What are those straight lines for? "XEmacs rules."
3:41 p.m.
-- "Stephen J. Turnbull" <turnbull(a)sk.tsukuba.ac.jp> spake thusly:
Michael> I do something similar:
Not really.
Michael> remove Control = Control_L
Michael> remove Lock = Caps_Lock
Michael> keysym Control_R = Caps_Lock
So far, so good. You now have no key that can do lock, and you have a
key in the lower right corner that does control. That key generates
the keysym Caps_Lock. The key in the lower left corner generates the
keysym Control_L, and the key to the left of the "A" generates the
keysym Caps_Lock.
Michael> add Control = Caps_Lock
You have added the control modifier to the key to the left of "A" and
(redundantly) to the key in the lower right corner.
Michael> keycode 37 = F22
I assume this keycode corresponds to the key in the lower left corner?
That key now generates the keysym F22. You have no Control keys
anymore, although you can generate control keystrokes. Summarizing:
Physical Keysym Modifier
------------------ --------- --------
Left of "A" Caps_Lock control
Lower left corner F22
Lower right corner Caps_Lock control
It might also help if someone were to point out the commands that need to be added to the
current configuration to get the appropiate behavior.
I'll attempt to do so, but someone please correct me if I'm wrong.
I know for a fact that you want to make "Left of 'A'" a control key. To
do this, figure out what the keycode of that key is (using xev), and then add to .xmodmap
(assuming your keyboard is like mine, which it probably is):
keycode 66 = Control_L
Then I'm going to make a leap and guess that you want the lower right corner to be a
Caps Lock. So you could do:
remove Control = Caps_Lock
add Lock = Caps_Lock
This should give you a configuration:
Physical Keysym Modifier
------------------ --------- --------
Left of "A" Control_L control
Lower left corner F22
Lower right corner Caps_Lock (caps) lock
With the set of commands:
remove Control = Control_L
remove Lock = Caps_Lock
keysym Control_R = Caps_Lock
add Control = Caps_Lock
keycode 37 = F22
keycode 66 = Control_L
remove Control = Caps_Lock
add Lock = Caps_Lock
I've tried launching XEmacs with this configuration, and it doesn't complain.
There's a probably a shorter way to get what you want (it seems like that
configuration does a bit too much moving things around), but I'll leave that as an
exercise for you if you want to pursue it.
Now I have to figure out how to fix my keyboard. ;-)
3:51 p.m.
Matt Tucker <tuck(a)whistlingfish.net> writes:
I've tried launching XEmacs with this configuration, and it
doesn't
complain.
It is also my experience that, when I do things right, XEmacs no
longer complains. I know because I set up caps lock to be control,
left windows key to be Alt, left Alt key to be meta, and the two
right-hand-side windows keys to be Super and Hyper.
As long as you resist the temptation to just add the appropriate
modifier key to the keysym, everything should work.
9:16 p.m.
>>>> On 19 Jan 2001 16:51:15 +0100,
>>>> In message <sxs1ytz5ymk.fsf(a)florida.arsdigita.de>,
>>>> Hrvoje Niksic <hniksic(a)arsdigita.com> wrote:
Hrvoje> As long as you resist the temptation to just add the
Hrvoje> appropriate modifier key to the keysym, everything should
Hrvoje> work.
Perhaps better stated as:
As long as you resist the temptation to just add the inappropriate
keysym to the modifier, everything should work.
i.e. if you want the key to be a "Foo" modifier, change the keysym of
the key to a "Foo" keysym before adding it to the "Foo" modifier.
1:55 a.m.
>>>> "Matt" == Matt Tucker
<tuck(a)whistlingfish.net> writes:
Matt> -- "Stephen J. Turnbull" <turnbull(a)sk.tsukuba.ac.jp> spake
thusly:
[...]
Matt> It might also help if someone were to point out the commands
Matt> that need to be added to the current configuration to get
Matt> the appropiate behavior.
The omission was deliberate. Getting it right is a job for Superman,
or a computer. Pay no attention to the man behind the curtain. Or in
front of it, for that matter.
Just get XKeyCaps.[1]
FAQ patch coming up.
Footnotes:
[1] Jamie did not pay me to write this.
--
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091
_________________ _________________ _________________ _________________
What are those straight lines for? "XEmacs rules."
2:41 p.m.
>>>> "SJT" == Stephen J Turnbull
<turnbull(a)sk.tsukuba.ac.jp> writes:
>>>> "Matt" == Matt Tucker
<tuck(a)whistlingfish.net> writes:
Matt> -- "Stephen J. Turnbull"
<turnbull(a)sk.tsukuba.ac.jp> spake thusly:
SJT> [...]
Matt> It might also help if someone were to point out the commands
Matt> that need to be added to the current configuration to get
Matt> the appropiate behavior.
SJT> The omission was deliberate. Getting it right is a job for Superman,
SJT> or a computer. Pay no attention to the man behind the curtain. Or in
SJT> front of it, for that matter.
It's possible to write a perl program in a few lines that will map the
key currently mapped to capslock, if there is one, and make it work
like control. It's harder to come up with a general solution that
will work for all keys. A big part of the problem is that I know of
no way to find out the ORIGINAL mapping for a key.
You'd like to have a program that would accept input like
mapkey caps_lock control
that will take care of modifier stuff and always keep the state of the
keyboard "legal" according to ICCCM. xmodmap is not such a program.
4:59 a.m.
I finally got around to play with the different suggestions.
The best seems to be a cocktail made out of the suggestions by Martin and Matt:
clear lock
clear control
keycode 66 = Control_L (66 is the key to the left of "A")
add Control = Control_L
keysym Control_R = Caps_Lock
add Lock = Caps_Lock
keycode 37 = F22
This makes caps lock (the key next to "A") into the control key, the left
control key into F22, and the right control key into the caps lock.
And XEmacs doesn't complain!
Still, if I try to simplify
keysym Control_R = Caps_Lock
add Lock = Caps_Lock
to
add Lock = Control_R
(as suggested by Martin) then I get a complaint that "Control_R does
locking, which is nonsensical". These two are essentially equivalent
if I understand 'man xmodmap' correctly (which I could never figure out
completely).
It seems that XEmacs assumes a specific function for every keysym name...
Thanks for all the suggestions.
--michael
Matt Tucker <tuck(a)whistlingfish.net> said:
-- "Stephen J. Turnbull" <turnbull(a)sk.tsukuba.ac.jp> spake thusly:
> Michael> I do something similar:
>
> Not really.
>
> Michael> remove Control =3D Control_L
> Michael> remove Lock =3D Caps_Lock
> Michael> keysym Control_R =3D Caps_Lock
>
> So far, so good. You now have no key that can do lock, and you have =
a
> key in the lower right corner that does control. That key generates
> the keysym Caps_Lock. The key in the lower left corner generates the
> keysym Control_L, and the key to the left of the "A" generates the
> keysym Caps_Lock.
>
> Michael> add Control =3D Caps_Lock
>
> You have added the control modifier to the key to the left of "A" and
> (redundantly) to the key in the lower right corner.
>
> Michael> keycode 37 =3D F22
>
> I assume this keycode corresponds to the key in the lower left =
corner?
> That key now generates the keysym F22. You have no Control keys
> anymore, although you can generate control keystrokes. Summarizing:
>
> Physical Keysym Modifier
> ------------------ --------- --------
> Left of "A" Caps_Lock control
> Lower left corner F22
> Lower right corner Caps_Lock control
It might also help if someone were to point out the commands that need =
to be added to the current configuration to get the appropiate =
behavior.
I'll attempt to do so, but someone please correct me if I'm wrong.
I know for a fact that you want to make "Left of 'A'" a control key. To
=
do this, figure out what the keycode of that key is (using xev), and =
then add to .xmodmap (assuming your keyboard is like mine, which it =
probably is):
keycode 66 =3D Control_L
Then I'm going to make a leap and guess that you want the lower right =
corner to be a Caps Lock. So you could do:
remove Control =3D Caps_Lock
add Lock =3D Caps_Lock
This should give you a configuration:
Physical Keysym Modifier
------------------ --------- --------
Left of "A" Control_L control
Lower left corner F22
Lower right corner Caps_Lock (caps) lock
With the set of commands:
remove Control =3D Control_L
remove Lock =3D Caps_Lock
keysym Control_R =3D Caps_Lock
add Control =3D Caps_Lock
keycode 37 =3D F22
keycode 66 =3D Control_L
remove Control =3D Caps_Lock
add Lock =3D Caps_Lock
I've tried launching XEmacs with this configuration, and it doesn't =
complain. There's a probably a shorter way to get what you want (it =
seems like that configuration does a bit too much moving things =
around), but I'll leave that as an exercise for you if you want to =
pursue it.
Now I have to figure out how to fix my keyboard. ;-)
8454
days inactive
8492
days old
13 comments
8 participants
participants (8)
-
Giacomo Boffi -
Hamish Macdonald -
Hrvoje Niksic -
Martin Buchholz -
Matt Tucker -
Michael Kifer -
Olivier Galibert -
Stephen J. Turnbull