On 5.02.08, rgheck wrote:
> G. Milde wrote:
> Nothing's being done immediately. Actually, it would help a lot if you
> could play around with the changes that have been suggested, namely:
My experiences with Debian Gnu/Linux and German keyboard:
> \bind "M-apostrophe" "quote-insert single"
> \bind "C-apostrophe" "accent-acute"
Works with acute-accent (which is unshifted [´`] key) but not with the
apostrophe (shifted [#'] key), as on my system, pressing either
acute-accent or apostrophe actually sends the apostrophe (') char.
> \bind "S-M-quotedbl" "self-insert \""
> \bind "S-C-quotedbl" "accent-umlaut"
Works here as expected (however it will fail on keyboards that have " in
an unshifted position!).
> As you may also have seen, we think there may be a bug here somewhere, too.
I suppose I found the bug:
>From http://wiki.lyx.org/Tips/KeyboardShortcuts
Precedence of keyboard bindings
-------------------------------
Note that the latest key binding takes precedence, ...
An exception are bindings using the modifier "~S", which doesnot overide
other bindings. ...
However, bindings containing "~S" do also not overwrite earlier bindings
with the same specifier, i.e.
\bind "S-M-quotedbl" "self-insert"
\bind "S-M-quotedbl" "accent-umlaut"
will bind quotedbl to self-insert (regardless of the Shift state).
As cua.sl contains a "~S"-binding for quotedbl:
\bind "~S-M-quotedbl" "quote-insert single"
the proposed bindings will not override this and hence not work.
The fix would (IMO) be
* let "~S"-bindings overrule earlier "~S"-bindings similar to "normal"
bindings. (But keep the precedence of more specific bindings.)
* official documentation of the keybindings syntax
This would result in clear and flexible rules for precedence:
1. Bindings overwrite earlier bindings with equal specifier, e.g.
\bind "~S-a" "foo"
\bind "~S-a" "bar"
will bind [A] and [a] to function "bar".
2. Specific bindings have higher precedence than more generic ones
regardless of position, e.g.
\bind "S-a" "foo"
\bind "~S-a" "bar"
will bind [A] to "foo" (but [a] to "bar").
It would also allow keyboard-layout-ignorant binding definitions of the
accent-* functions.
OTOH, you can of course use the "~S" form already now if you uncomment or
remove conflicting earlier definitions...
BTW, my private gm.bind contains:
# accents (dead keys)
\bind "M-~S-asciicircum" "accent-circumflex" # also "M-m h"
\bind "M-~S-asciitilde" "accent-tilde" # also "M-m asciitilde"
\bind "M-~S-degree" "accent-circle" # °
\bind "M-~S-plus" "accent-tilde" # ~ (AltGr-+ on German keyb.)
\bind "M-~S-quoteleft" "accent-grave" # `
\bind "M-~S-apostrophe" "accent-acute" # '
\bind "M-v" "accent-caron" # v = hacheck
Günter
