From storm@olicom.dk Thu Dec 12 00:48:07 1996 X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] [nil "Thu" "12" "December" "1996" "09:47:45" "+0100" "Kim F. Storm" "storm@olicom.dk" nil "208" "Suggestion for a \"transient-mark-map\" in Emacs (repost)" "^From:" nil nil "12" nil nil nil nil] nil) Received: from olicom.olicom.dk (eolicom.olicom.dk [194.255.19.1]) by june.cs.washington.edu (8.8.3+CSE/7.2ju) with ESMTP id AAA11115 for ; Thu, 12 Dec 1996 00:48:03 -0800 Received: from kfs.olicom.dk (kfs.olicom.dk [172.16.99.27]) by olicom.olicom.dk (8.7.5/8.7.3/efh.2.20) with SMTP id JAA12156; Thu, 12 Dec 1996 09:47:45 +0100 (MET) Message-Id: <199612120847.JAA12156@olicom.olicom.dk> In-reply-to: <199612112013.PAA20665@psilocin.gnu.ai.mit.edu> (message from Richard Stallman on Wed, 11 Dec 1996 15:13:08 -0500) From: "Kim F. Storm" To: rms@gnu.ai.mit.edu Cc: andrewi@harlequin.co.uk, voelker@cs.washington.edu Subject: Suggestion for a "transient-mark-map" in Emacs (repost) Date: Thu, 12 Dec 1996 09:47:45 +0100 (MET) It seems that the following message was lost is transit. >From: "Kim F. Storm" >To: andrewi@harlequin.co.uk, voelker@cs.washington.edu, rms@gnu.ai.mit.edu >Date: Mon Dec 09 12:13:07 +0000 1996 >Subject: Suggestion for a "transient-mark-map" in Emacs I have recently started using the pc-select.el style of marking and selecting text in NT-emacs (on Windows 95), and I actually find it very convenient. However, since I switch between emacs and other Windows 95 applications, I frequently made the mistake of using C-x and C-c for cut and copy when I'm in emacs. I tried to find a way to map C-x and C-c to kill-region and copy-region-as-kill when the region mark was active and keep their normal prefix functionality otherwise. However, I couldn't find an easy way to do it, so instead I came up with the idea of having a special keymap which was only used when the mark was active and the transient-mark-mode is selected; I called this the transient-mark-map. Now, with this new map, it is straightforward to map C-x and C-c as kill and copy when the region is active. My experience with this temporary remapping of C-x and C-c is quite positive; I do make the mistake of deleting a random region with C-x from time to time, but it is easy to undo (I also remapped C-z to undo and C-v to paste). In any case, this happens much less frequently than before when I often pressed C-x (in vain) to kill the region. Below, I supply my diffs to keyboard.c and keymap.c (relative to the 19.34 release). Mon Dec 09 11:55:00 1996 Kim Storm * keymap.c (Vtransient_mark_map): new variable. (key-binding): Lookup key in Vtransient_mark_map. (syms_of_keymap): Defvar transient-mark-map. * keyboard.c (read_key_sequence): Allow more dynamic construction of the list of active submaps. Added Vtransient_mark_map to submaps when buffer's mark is active in transient-mark-mode. *** keymap.org Tue Jun 11 02:58:18 1996 --- keymap.c Fri Nov 29 11:01:34 1996 *************** *** 78,83 **** --- 78,86 ---- /* Keymap mapping ASCII function key sequences onto their preferred forms. */ Lisp_Object Vkey_translation_map; + /* The keymap used before any other map if transient mark is active */ + Lisp_Object Vtransient_mark_map; + /* A list of all commands given new bindings since a certain time when nil was stored here. This is used to speed up recomputation of menu key equivalents *************** *** 1086,1091 **** --- 1089,1102 ---- { Lisp_Object local; + if (! NILP (Vtransient_mark_map) + && ! NILP (Vtransient_mark_mode) && !NILP (current_buffer->mark_active)) + { + value = Flookup_key (Vtransient_mark_map, key, accept_default); + if (! NILP (value) && !INTEGERP (value)) + RETURN_UNGCPRO (value); + } + nmaps = current_minor_maps (0, &maps); /* Note that all these maps are GCPRO'd in the places where we found them. */ *************** *** 2749,2754 **** --- 2760,2769 ---- DEFVAR_LISP ("minibuffer-local-must-match-map", &Vminibuffer_local_must_match_map, "Local keymap for minibuffer input with completion, for exact match."); Vminibuffer_local_must_match_map = Fmake_sparse_keymap (Qnil); + + DEFVAR_LISP ("transient-mark-map", &Vtransient_mark_map, + "Keymap to use when transient-mark mode is active in the current buffer."); + Vtransient_mark_map = Fmake_sparse_keymap (Qnil); current_global_map = global_map; *** keyboard.org Fri Aug 23 09:56:00 1996 --- keyboard.c Fri Nov 29 11:41:14 1996 *************** *** 333,338 **** --- 333,341 ---- This one takes precedence over ordinary definitions. */ extern Lisp_Object Vkey_translation_map; + /* The keymap used before any other map if transient mark is active */ + extern Lisp_Object Vtransient_mark_map; + /* Non-nil means deactivate the mark at end of this command. */ Lisp_Object Vdeactivate_mark; *************** *** 5770,5775 **** --- 5773,5779 ---- keymaps we have allocated space for. */ int nmaps; int nmaps_allocated = 0; + int ndefs_allocated = 0; /* defs[0..nmaps-1] are the definitions of KEYBUF[0..t-1] in the current keymaps. */ *************** *** 5914,5931 **** click and we need to switch buffers, we jump back here to rebuild the initial keymaps from the current buffer. */ { ! Lisp_Object *maps; if (!NILP (current_kboard->Voverriding_terminal_local_map) || !NILP (Voverriding_local_map)) { - if (3 > nmaps_allocated) - { - submaps = (Lisp_Object *) alloca (3 * sizeof (submaps[0])); - defs = (Lisp_Object *) alloca (3 * sizeof (defs[0])); - nmaps_allocated = 3; - } - nmaps = 0; if (!NILP (current_kboard->Voverriding_terminal_local_map)) submaps[nmaps++] = current_kboard->Voverriding_terminal_local_map; if (!NILP (Voverriding_local_map)) --- 5918,5939 ---- click and we need to switch buffers, we jump back here to rebuild the initial keymaps from the current buffer. */ { ! if (10 > nmaps_allocated) ! { ! submaps = (Lisp_Object *) alloca (10 * sizeof (submaps[0])); ! nmaps_allocated = 10; ! } ! nmaps = 0; ! ! if (!NILP (Vtransient_mark_map) ! && !NILP (Vtransient_mark_mode) && !NILP (current_buffer->mark_active)) ! { ! submaps[nmaps++] = Vtransient_mark_map; ! } if (!NILP (current_kboard->Voverriding_terminal_local_map) || !NILP (Voverriding_local_map)) { if (!NILP (current_kboard->Voverriding_terminal_local_map)) submaps[nmaps++] = current_kboard->Voverriding_terminal_local_map; if (!NILP (Voverriding_local_map)) *************** *** 5933,5946 **** } else { ! nmaps = current_minor_maps (0, &maps); ! if (nmaps + 2 > nmaps_allocated) { ! submaps = (Lisp_Object *) alloca ((nmaps+2) * sizeof (submaps[0])); ! defs = (Lisp_Object *) alloca ((nmaps+2) * sizeof (defs[0])); ! nmaps_allocated = nmaps + 2; } ! bcopy (maps, submaps, nmaps * sizeof (submaps[0])); #ifdef USE_TEXT_PROPERTIES submaps[nmaps++] = orig_local_map; #else --- 5941,5959 ---- } else { ! Lisp_Object *maps; ! int mi_maps; ! ! mi_maps = current_minor_maps (0, &maps); ! if (mi_maps + nmaps + 2 > nmaps_allocated) { ! Lisp_Object *omaps = submaps; ! nmaps_allocated = mi_maps + nmaps + 2; ! submaps = (Lisp_Object *) alloca (nmaps_allocated * sizeof (submaps[0])); ! bcopy (omaps, submaps, nmaps * sizeof (submaps[0])); } ! bcopy (maps, submaps + nmaps, mi_maps * sizeof (submaps[0])); ! nmaps += mi_maps; #ifdef USE_TEXT_PROPERTIES submaps[nmaps++] = orig_local_map; #else *************** *** 5949,5954 **** --- 5962,5973 ---- } submaps[nmaps++] = current_global_map; } + + if (ndefs_allocated < nmaps_allocated) + { + ndefs_allocated = nmaps_allocated; + defs = (Lisp_Object *) alloca (ndefs_allocated * sizeof (defs[0])); + } /* Find an accurate initial value for first_binding. */ for (first_binding = 0; first_binding < nmaps; first_binding++) -- Kim F. Storm, Olicom A/S From andrewi@harlequin.co.uk Thu Dec 12 02:58:47 1996 X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] [nil "Thu" "12" "December" "1996" "10:58:14" "GMT" "Andrew Innes" "andrewi@harlequin.co.uk" nil "32" "Re: Suggestion for a \"transient-mark-map\" in Emacs (repost)" "^From:" nil nil "12" nil nil nil nil] nil) Received: from holly.cam.harlequin.co.uk (holly.cam.harlequin.co.uk [193.128.4.58]) by june.cs.washington.edu (8.8.3+CSE/7.2ju) with SMTP id CAA15996 for ; Thu, 12 Dec 1996 02:58:46 -0800 Received: from propos.long.harlequin.co.uk by holly.cam.harlequin.co.uk; Thu, 12 Dec 1996 10:58:21 GMT Received: from elan.long.harlequin.co.uk (elan [193.128.93.78]) by propos.long.harlequin.co.uk (8.6.12/8.6.12) with SMTP id KAA01063; Thu, 12 Dec 1996 10:58:14 GMT Message-Id: <199612121058.KAA01063@propos.long.harlequin.co.uk> In-Reply-To: <199612120847.JAA12156@olicom.olicom.dk> (storm@olicom.dk) From: Andrew Innes To: storm@olicom.dk Cc: rms@gnu.ai.mit.edu, voelker@cs.washington.edu Subject: Re: Suggestion for a "transient-mark-map" in Emacs (repost) Date: Thu, 12 Dec 1996 10:58:14 GMT On Thu, 12 Dec 1996 09:47:45 +0100 (MET), "Kim F. Storm" said: > >It seems that the following message was lost is transit. I received the first copy, but didn't have any real comment to make. >>From: "Kim F. Storm" >>To: andrewi@harlequin.co.uk, voelker@cs.washington.edu, rms@gnu.ai.mit.edu >>Date: Mon Dec 09 12:13:07 +0000 1996 >>Subject: Suggestion for a "transient-mark-map" in Emacs > >I have recently started using the pc-select.el style of marking and >selecting text in NT-emacs (on Windows 95), and I actually find it >very convenient. However, since I switch between emacs and other >Windows 95 applications, I frequently made the mistake of using C-x >and C-c for cut and copy when I'm in emacs. > >I tried to find a way to map C-x and C-c to kill-region and >copy-region-as-kill when the region mark was active and keep their >normal prefix functionality otherwise. However, I couldn't find an >easy way to do it, so instead I came up with the idea of having a >special keymap which was only used when the mark was active and the >transient-mark-mode is selected; I called this the transient-mark-map. This sounds okay to me, but then I don't use transient-mark-mode. I gather your change wouldn't affect me in that case, so that's okay. Are you proposing to alter pc-select.el as well to define C-x, C-v, C-c and C-z as standard in transient-mark-map, perhaps as a configurable option? AndrewI From storm@olicom.dk Thu Dec 12 07:12:54 1996 X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] [nil "Thu" "12" "December" "1996" "16:12:43" "+0100" "Kim F. Storm" "storm@olicom.dk" nil "39" "Re: Suggestion for a \"transient-mark-map\" in Emacs (repost)" "^From:" nil nil "12" nil nil nil nil] nil) Received: from olicom.olicom.dk (eolicom.olicom.dk [194.255.19.1]) by june.cs.washington.edu (8.8.3+CSE/7.2ju) with ESMTP id HAA26283 for ; Thu, 12 Dec 1996 07:12:52 -0800 Received: from kfs.olicom.dk (kfs.olicom.dk [172.16.99.27]) by olicom.olicom.dk (8.7.5/8.7.3/efh.2.20) with SMTP id QAA03884; Thu, 12 Dec 1996 16:12:43 +0100 (MET) Message-Id: <199612121512.QAA03884@olicom.olicom.dk> In-reply-to: <199612121058.KAA01063@propos.long.harlequin.co.uk> (message from Andrew Innes on Thu, 12 Dec 1996 10:58:14 GMT) From: "Kim F. Storm" To: andrewi@harlequin.co.uk CC: rms@gnu.ai.mit.edu, voelker@cs.washington.edu Subject: Re: Suggestion for a "transient-mark-map" in Emacs (repost) Date: Thu, 12 Dec 1996 16:12:43 +0100 (MET) Date: Thu, 12 Dec 1996 10:58:14 GMT From: Andrew Innes >>From: "Kim F. Storm" > >I have recently started using the pc-select.el style of marking and >selecting text in NT-emacs (on Windows 95), and I actually find it >very convenient. However, since I switch between emacs and other >Windows 95 applications, I frequently made the mistake of using C-x >and C-c for cut and copy when I'm in emacs. > >I tried to find a way to map C-x and C-c to kill-region and >copy-region-as-kill when the region mark was active and keep their >normal prefix functionality otherwise. However, I couldn't find an >easy way to do it, so instead I came up with the idea of having a >special keymap which was only used when the mark was active and the >transient-mark-mode is selected; I called this the transient-mark-map. This sounds okay to me, but then I don't use transient-mark-mode. I gather your change wouldn't affect me in that case, so that's okay. And since the transient-mark-map is empty by default, it shouldn't affect anybody else for that matter. Are you proposing to alter pc-select.el as well to define C-x, C-v, C-c and C-z as standard in transient-mark-map, perhaps as a configurable option? Yes, although C-z and C-v must be generally remapped in global-map since they are not linked to the active region as such. I have made some steps to improve on pc-selection-mode and delete-selection-mode, but I haven't had the time to stream-line the code. In any case, my changes to pc-select.el are not mandatory to take advantage of the transient-mark-map, provided you remap the keys yourself. -- Kim F. Storm, Olicom A/S From storm@olicom.dk Fri Dec 13 01:59:27 1996 X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] [nil "Fri" "13" "December" "1996" "10:59:01" "+0100" "Kim F. Storm" "storm@olicom.dk" nil "41" "Re: Suggestion for a \"transient-mark-map\" in Emacs (repost)" "^From:" nil nil "12" nil nil nil nil] nil) Received: from olicom.olicom.dk (eolicom.olicom.dk [194.255.19.1]) by june.cs.washington.edu (8.8.3+CSE/7.2ju) with ESMTP id BAA11158 for ; Fri, 13 Dec 1996 01:59:24 -0800 Received: from kfs.olicom.dk (kfs.olicom.dk [172.16.99.27]) by olicom.olicom.dk (8.7.5/8.7.3/efh.2.20) with SMTP id KAA27750; Fri, 13 Dec 1996 10:59:01 +0100 (MET) Message-Id: <199612130959.KAA27750@olicom.olicom.dk> In-reply-to: <199612130129.UAA02823@psilocin.gnu.ai.mit.edu> (message from Richard Stallman on Thu, 12 Dec 1996 20:29:55 -0500) From: "Kim F. Storm" To: rms@gnu.ai.mit.edu CC: andrewi@harlequin.co.uk, voelker@cs.washington.edu Subject: Re: Suggestion for a "transient-mark-map" in Emacs (repost) Date: Fri, 13 Dec 1996 10:59:01 +0100 (MET) Date: Thu, 12 Dec 1996 20:29:55 -0500 From: Richard Stallman Cc: andrewi@harlequin.co.uk, voelker@cs.washington.edu I do not mind if you remap C-x, C-v, C-c and C-z, but I do not want to install such a change in Emacs. Sorry. Maybe I didn't make myself clear, but I didn't ask you to remap anything, I just asked you to make it possible for me to do it. As it is now, I just cannot see any simple way to have C-x and C-c work on the region the way it does in all other applications I use on this system (Windows95). I cannot imagine how the proposed change (transient-mark-map) can harm anyone if they don't use it. The changes are small and in my opinion trivial. I honestly fail to see the reason why you object to include these changes. For the record, I've been maintaining emacs at this company for the last 5 years, and before that at other sites for another 5 years, and have always advocated its use (over the official company standard which is the "brief" editor, its successor "crisp", or the various MS editors). The main reason of course because it is a superb editor, but also because it has two major features none of the other editors have: - you have access to the source code, and if something doesn't work (the way you want it to), you fix it, and - you share the work you do to improve emacs - and the work a lot of other users have done - to achieve an even more superior product. Of course, I can understand that you cannot simply put each and every proposal into the official code, but given the amount and varity of features that is already in there, I find it hard to distinguish what is acceptable and what is not, and why my proposal fall into the last category. So I'm sorry too. -- Kim F. Storm, Olicom A/S From rms@gnu.ai.mit.edu Thu Dec 12 17:29:28 1996 X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] [nil "Thu" "12" "December" "1996" "20:29:55" "-0500" "Richard Stallman" "rms@gnu.ai.mit.edu" nil "6" "Re: Suggestion for a \"transient-mark-map\" in Emacs (repost)" "^From:" nil nil "12" nil nil nil nil] nil) Received: from psilocin.gnu.ai.mit.edu (psilocin.gnu.ai.mit.edu [128.52.46.62]) by june.cs.washington.edu (8.8.3+CSE/7.2ju) with SMTP id RAA14515 for ; Thu, 12 Dec 1996 17:29:27 -0800 Received: by psilocin.gnu.ai.mit.edu (8.6.12/8.6.12GNU) id UAA02823; Thu, 12 Dec 1996 20:29:55 -0500 Message-Id: <199612130129.UAA02823@psilocin.gnu.ai.mit.edu> In-reply-to: <199612121512.QAA03884@olicom.olicom.dk> (storm@olicom.dk) From: Richard Stallman To: storm@olicom.dk CC: andrewi@harlequin.co.uk, voelker@cs.washington.edu Subject: Re: Suggestion for a "transient-mark-map" in Emacs (repost) Date: Thu, 12 Dec 1996 20:29:55 -0500 I do not mind if you remap C-x, C-v, C-c and C-z, but I do not want to install such a change in Emacs. Sorry.