Eli Zaretskii <[email protected]> writes: >> From: Kenichi Handa <[email protected]> >> Cc: [email protected], [email protected] >> Date: Mon, 28 Jun 2010 15:02:30 +0900 >> >> > I don't think so. I think the Emacs translates a character regardless >> > of which key produced it. Take a look at leim/quail/hebrew.el: all it >> > tells Emacs is the characters to translate. There's no information >> > there regarding the layout of the keyboard. >> >> Sure. All input methods should be written assuming >> `quail-keyboard-layout-standard'. Have you seen the help >> of `hebrew' input method? It says: >> >> ------------------------------------------------------------ >> ... >> This input method works by translating individual input characters. >> Assuming that your actual keyboard has the `standard' layout, >> translation results in the following "virtual" keyboard layout: > > How does quail use the information about the layout of the keyboard? > All I see in the definition of the input method is a list of > characters and their translations. Why should it matter which > keyboard key produced a character? Assuming that the key that > produces `a' in English should always produce SHIN in Hebrew, why > should it matter what is the number of that key on the keyboard and > what is its location?
I cannot say, philosophically, whether this is the right things to do or not (although my instincts are to say that is is). I can, however, repeat the comment from quail.el: ;; Keyboard layout translation handlers. ;; Some Quail packages provide localized keyboard simulation which ;; requires a particular keyboard layout. In this case, what we need ;; is locations of keys the user entered, not character codes ;; generated by those keys. However, for the moment, there's no ;; common way to get such information. So, we ask a user to give ;; information of his own keyboard layout, then translate it to the ;; standard layout which we defined so that all Quail packages depend ;; just on it. -- Michael Welsh Duggan ([email protected]) _______________________________________________ emacs-bidi mailing list [email protected] http://lists.gnu.org/mailman/listinfo/emacs-bidi
