|Summary:||home/end/backspace/del key under xterm/rxvt|
|Product:||[Retired] Red Hat Linux||Reporter:||richb|
|Component:||distribution||Assignee:||Preston Brown <pbrown>|
|Status:||CLOSED CURRENTRELEASE||QA Contact:|
|Fixed In Version:||Doc Type:||Bug Fix|
|Doc Text:||Story Points:||---|
|Last Closed:||1999-03-25 19:40:28 UTC||Type:||---|
|oVirt Team:||---||RHEL 7.3 requirements from Atomic Host:|
|Cloudforms Team:||---||Target Upstream Version:|
Description richb 1998-11-09 21:08:25 UTC
I'm writing to try and address the issue of the home/end/del/backspace keys in xterm and rxvt. Out of the box these keys don't work correctly with RedHat Linux. It turns out the issues are quite involved: Here's a summary of my findings: 1/ The 'nxterm' shipped with RedHat seems to be a branch of an old version of xterm that contains a bug wrt to the home/end keys. 2/ The 'xterm' termcap/terminfo entries for home/end do not match the actual codes generated by xterm. 3/ The 'xterm' termcap/terminfo entries for del/bs do not match the actual codes generated by rxvt. 4/ The termcap/terminfo entries for 'xterm' contain non-standard codes, which I believe is undesirable. 5/ Given that xterm and rxvt generate different codes for these keys, it's necessary to either: a) use app-default/Xterm or app-defaults/Rxvt to map one program's keys onto the other, or b) Get rxvt to set TERM=rxvt rather than TERM=xterm. I believe that finding the best solution requires some discussion. Could someone from RedHat please contact me so I can offer my proposal. Hoping you are listening... Richard Browne.
Comment 1 Cristian Gafton 1998-12-03 21:23:59 UTC
1. what is the nature of the bug in nxterm ? 2. is there a patch available ? 3. is there a patch available ? 4. Which codes are non standard ? 5. 'rxvt' is not a "common" terminal name and it will interract badly when Linux is used in in a mixed environment, with other unices around. I'd rather make rxvt match the xterm keys, altough I am not sure how many apps that already know about rxvt will be broken by this...
Comment 2 richb 1998-12-08 13:21:59 UTC
1. According to the current xterm maintainer, Thomas E. Dickey (email@example.com), older versions of xterm had an array indexing bug. The home/end keys generated garbage. nxterm seems to be a fork of an old xterm and therefore has this bug. In the case of the nxterm binaries shipped with RedHat 5.1, home generates \E[^@ and end generates \E[e. Nobody has been able to tell me exactly what nxterm does. I suspect it may be obsolete. Certainly color_xterm (another variant, possibly related to nxterm) is obsolete since xterm itself has had color capabilities for some time. 2. This bug was fixed in xterm for the XFree86 distribution in 1996. The values xterm generates for home/end are \EOH and \EOF. My proposal is this: 1. Accept that the 'correct' codes for home/end/backspace/del are those generated by xterm: \EOH, \EOF, ^H, and \E[3~. 2. Anything that claims to be xterm compatible should generate the same codes as xterm. I can't understand the rationale for rxvt setting TERM=xterm and then generating different keycodes, it's crazy. 3. Reset the xterm termcap/terminfo entries to the 'correct' values. The values being obtained from the XFree86 distribution, or from the master terminfo sources currently maintained by Eric S Raymond. 4. Consider scrapping nxterm and color_xterm. What do they do? If they must be kept, try app-defaults/NXTerm to get correct values. I'm unsure if the array indexing bug can be worked around like this, I haven't tried it. 5. Either: a) Make rxvt set TERM=rxvt, then it can generate whatever codes it likes, or b) Use app-defaults/Rxvt to get it to generate xterm compatible codes. I know many people prefer the TERM=rxvt solution, but I'm not sure I want to comment. Either solution would work for me. 7. Finally, lose the default mapping that makes PgUp/PgDn scroll the xterm window. It's a pain in the bum when using full screen apps such as Midnight Commander. I have observed that this is particularly confusing for new users. If you would like me to do any further research I'd be happy to do so.
Comment 3 Preston Brown 1999-03-24 21:28:59 UTC
I'm working on this.
Comment 4 Preston Brown 1999-03-25 19:40:59 UTC
fixed in ncurses-4.2-17 and later.