Received: (at submit) by bugs.debian.org; 7 Mar 1997 05:12:16 +0000 Received: (qmail 8454 invoked from network); 7 Mar 1997 05:12:14 -0000 Received: from g62-99.citenet.net (brnstnd@kramden.acf.nyu.edu@207.183.46.99) by master.debian.org with SMTP; 7 Mar 1997 05:12:14 -0000 Received: by g62-99.citenet.net id m0w2rou-0008gkC (Debian Smail-3.2 1996-Jul-4 #2); Fri, 7 Mar 1997 00:03:52 -0500 (EST) Message-Id: Date: Fri, 7 Mar 1997 00:03:52 -0500 (EST) From: Greg Stark To: submit@bugs.debian.org X-Debian-CC: bug-gnu-emacs@prep.ai.mit.edu, report@XFree86.org, gsstark@mit.edu Subject: X key binding of shift tab on XFree86 3.2 (ISO_Left_Tab) Package: xbase Version: 3.2-3 In GNU Emacs 19.34.1 (i386-debian-linux-gnu, X toolkit) of Wed Aug 28 1996 on maneki-neko configured using `configure --prefix=/usr --with-pop=yes --with-x=yes --with-x-toolkit=lucid i386-debian-linux' On Xfree86 3.2 (actually 3.2A with some patches, but i'm pretty sure it isn't relevant) seems to bind tab by default such that shift-tab generates "ISO_Left_Tab" which i've never heard of, but presumably emacs should interpret that as the emacs key binding "backtab" or perhaps just ignore it and handle it as shift-tab. I'm partial to the former, but then i suggested emacs should map shift-tab to backtab if it's bound in any case, which you didn't like. It also seems like XLookupString should probably generate a TAB character so that programs can notice that it's a tab character with the shift modifier pressed but i'm not really sure if that's the right approach. Here's some relevant output from xmodmap -pk and xev: 23 0xff09 (Tab) 0xfe20 (ISO_Left_Tab) KeyPress event, serial 21, synthetic NO, window 0x1c00001, root 0x2b, subw 0x0, time 3006375125, (129,32), root:(210,113), state 0x1, keycode 23 (keysym 0xfe20, ISO_Left_Tab), same_screen YES, XLookupString gives 0 characters: ""