Josip Gracin wrote: Each modifier bit has to be associated with at most one modifier (Alt, Meta, Hyper, Super). You have mod1 associated with both Alt and Meta, and mod4 associated with both Super and Hyper. This won't work, as XEmacs cannot figure out which modifiers the mod1 and mod3 bits are supposed to represent. It normally generates a warning about this situation. You also need to take into account the fact that some toolkits (e.g. GTK) blindly treat mod1 as "the Alt/Meta modifier", while Xt and XEmacs handle modifier bits according to which keysyms are bound to them (as dictated by the ICCCM). Note that you cannot have all six of the Alt, Meta, Hyper, Super, Num_Lock and Mode_switch/ISO_Level3_Shift modifiers; there simply aren't enough bits (the Shift, Control and Lock modifiers are restricted to specific keysyms: Shift_L/R, Control_L/R and either Shift_Lock or Caps_Lock; binding them to other keysyms won't work). -- Glynn Clements <glynn(a)gclements.plus.com&gt; _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Josip Gracin wrote: Probably. If pressing Alt+X brings up a M-x prompt, then it's treating mod1 as Meta. I have both: shift Shift_L (0x32), Shift_R (0x3e) lock Caps_Lock (0x42) control Control_L (0x25), Control_R (0x6d) mod1 Alt_L (0x73), Alt_L (0x7f) mod2 Num_Lock (0x4d) mod3 mod4 Meta_L (0x40), Meta_L (0x7d), Meta_L (0x9c) mod5 Mode_switch (0x5d), ISO_Level3_Shift (0x7c) [From the keycodes, you can see that the "Windows" keys are Alt_* while the "Alt" keys are Meta_*.] If I press Windows+X (i.e. Alt-x), I get "A-x not defined." in the echo area; Alt+X (i.e. Meta-x) gives me the M-x prompt. I have Alt rather than Meta on mod1 so that I can use Alt-<key> for WM shortcuts without shadowing any of XEmacs' Meta-<key> bindings. E.g. Alt-Tab switches between windows, while Meta-Tab does completion within XEmacs. -- Glynn Clements <glynn(a)gclements.plus.com&gt; _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Josip Gracin <gracin <at> tel.fer.hr> wrote: Earlier in thread, Glynn Clements wrote: These are very useful functions, but their key bindings are difficult for KDE and Gnome users to press. Perhaps it'd be a worthwhile idea to create additional bindings that do these things? (In case you didn't know: M-TAB calls different functions depending what major mode is enabled. For example, in Text mode, it runs, IIRC, ispell-complete-word. This looks in the English dictionary and to see what word you are currently in the middle of typing, then helps you to complete it.) Regards, Jason _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Not for me (as a KDE user) at least. Kwin, the KDE window manager, can be configured not to interfere with those shortcuts. I'm heavily using META-something under KDE. Regards, Wulf _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

2007/3/13, Stephen J. Turnbull <stephen(a)xemacs.org&gt; writes: KDE and GNOME both bind M-TAB to something else (an application switcher) so the keystroke doesn't get through to Emacs. Control-Alt-<backspace> is awkward for all X Window System users. It is the hotkey for suddenly killing any running X server without providing any warning dialog. That's confusingly similar to C-M-<backspace>. Plus, on many machines, the same physical key produces both Alt and Meta. Can anyone suggest some bindings? I don't know the Emacs keymap so well. Hopeless ideas: Maybe we can use something with TAB in it instead of M-TAB? How about Shift-Tab? And for backward-kill-sexp, how about Shift-<backspace>? (But maybe those wouldn't work when running in console mode.) Time permitting, I will test any patches that address this issue on my Ubuntu 6.10 + Gnome PC. -- Jason Spiro: computer consulting with a smile. I also provide training and spyware removal services for homes and businesses. Call or email for a FREE 5-minute consultation. Satisfaction guaranteed. +1 (416) 781-5938 / Email: info(a)jspiro.com / MSN: jasonspiro(a)hotmail.com _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Jason Spiro wrote: If you have trouble with Alt/Meta-modified keypresses, you can always use the Escape prefix instead. Or you can use key-translation-map to globally map some other sequence to Meta-Tab, e.g.: (define-key key-translation-map [(control ?c) ?t] [(meta tab)]) The main problem with defining another key sequence is that they're essentially all taken. Individual users can forego a sequence that they don't themselves use, but the XEmacs project can't realistically impose such a change on all users. -- Glynn Clements <glynn(a)gclements.plus.com&gt; _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Glynn Clements writes: Argh. Is this a bug or not? I think this is what I would expect from the documentation and the fact that it's actually a keymap, but I'm not sure it's ever useful. _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Josip Gracin writes: No, XEmacs is not grokking it; the mapping is broken (definitely if XKB is not available, and as far as I know XKB-enabled servers are required to be backward compatible by the XKB spec) and XEmacs failed to behave as you expected as a result. It is possible that either XEmacs's keyboard handler has been patched or the warnings suppressed by the distro. (Mandrake, at least, deliberately broke XEmacs's behavior with respect to Alt/Meta in order to force compatibility with their broken scheme.) My guess is that the reason that this sort of works is that the context where Alt and Meta are used is applications, and they usually only have one of the two modifiers, and just look for their favorite modifier. Since both are on the mod1 bit, the user sees correct behavior. Similarly for Super and Hyper, which are mostly used to keep window manager key bindings out of the way of the applications. With the modern XKB extension, there may be other ways to handle modifiers. However, porting XEmacs to XKB is a fairly large task, compounded by the fact that the XKB developers provided no API documentation worthy of the name the last time I looked. I guess I'll have to take another look. Hard to say, without looking at exactly the configuration of XEmacs as well as the patches applied. What actually is supposed to happen is that XEmacs looks at the keyboard mapping, and checks it for sanity. At this point the "your X server keyboard configuration is broken" warnings are issued. Next, XEmacs checks the modifer map for a modifier bit assigned to a Meta key. If found, Meta (physical) is Meta (XEmacs). If not found, XEmacs checks for a modifier bit assigned to an Alt key. If found, Alt (physical) is Meta (XEmacs). If not found, there will be no Meta modifier, and you must use the ESC key as a prefix key to generate XEmacs Meta keystrokes. The reason for this behavior is that most historic Unix workstations had a Meta key. When PC-based X servers started to become popular, vendors had only the Alt key, and they had a choice: to assign it the Alt keysym or the Meta keysym. Most chose the former, reasonable as only the apps need to deal with the issue. Documenting the Alt-is- really-Meta choice would be hard, and many users would fail to notice and simply report a bug. I certainly couldn't blame them. _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta