after trying to get page-completions set to off the only way i could do this was setting it in /etc/inputrc according to the man page tho i should of been able to set via ~/.inputrc and i would of prefered that also i seem to recall seeing a similiar problem a few months ago with this also but didnt think much of it using current bash for x86 (2.05b-r7)
The problem is that /etc/profile sets INPUTRC=/etc/inputrc if you don't have a ~/.inputrc when you login. There's a few ways you can fix this: - override in .bashrc for new terminals - unset INPUTRC and exec bash - use login shells for terminals - log out and log back in to pick up the new INPUTRC The best way to fix this is for bash to use /etc/inputrc automatically as a last resort following ~/.inputrc. So... bash-3.0-r3 (and new readline should also follow with this functionality) Additionally I'll remove the INPUTRC setting from baselayout's /etc/profile in the next release of baselayout
Re-opening so that vapier can apply the patch to readline-5 app-shells/bash/files/bash-3.0-etc-inputrc.patch
By removing INPUTRC=/etc/inputrc peoples delete/page up/down etc cease to work as for the ones who don't have their own .inputrc file, /etc/inputrc isn't used. See http://forums.gentoo.org/viewtopic.php?t=171783 for example about problems. Maybe add .inputrc to /etc/skel with the content "$include /etc/inputrc"? So that at least new users will get it working. I of course would keep the environment variable in tact so that it would work for non-advanced users... The ones who want to tune it, could add "export INPUTRC=~/.inputrc" to their ~/.bashrc file. Or maybe even add "#export INPUTRC=~/.inputrc" as a hint to /etc/skel/.bashrc Just my two cents about the matter that has proved to give headache to some people, including me until I had time to look into it. I really missed my delete (under cursor) key ;)
Just ran into the same problem, don't use the PageUp/Down but the Delete key I do! Doesn't appear that the bash-3.0-etc-inputrc.patch is working completely or I haven't fully comprehened what it is doing. Have added: ]if [ -z "$INPUTRC" -a ! -f "$HOME/.inputrc" ] then export INPUTRC="/etc/inputrc" fi back into /etc/profile on my system so have Del key back. So is the patch doing it's job or is there a better way to do this?
So is this a bug or "feature"? Presumably these key lines were not removed from /etc/profile by accident. If it was intentional, why? This sort unnecessary change to the way the system behaves will quickly make maintaining a reliable system a full time job. I cant calculate how many hours it has cost me to find out were this disfunctionality came from. Please can we have consistancy unless the is a major over-riding importance to make such changes.
> quickly make maintaining a reliable system a full time job You're running ~arch if you got the updated profile. That's no way to maintain a reliable system. This stuff works fine with ~arch bash-3.0-r6 (with no INPUTRC in the environment). I just tested and put extra commands in /etc/inputrc just to make sure. If other things are broken it's because readline still needs to be updated.
I'm not running ~arch.
readline-5.0-r1 has the /etc/inputrc patch, thanks all
Don't know what is wrong here then. Still have to put if [ -z "$INPUTRC" -a ! -f "$HOME/.inputrc" ]; then export INPUTRC="/etc/inputrc" fi in /etc/profile or delete key give ~ and don't work. Have readline-5.0-r1 . Guess it is just a bug going to have to live with when etc-update.