When I start MC as <user>, outside of X in a VT, I cannot navigate up/down the panels using the cursor keys, and the function keys do not work either. Basically I can't use MC as <user>

If I login or su or su - to root, it works as it should be, everything is fine. But, if I su or su - to <user> from root, it again doesn't work properly.

I'm not sure if this is something I've done recently that screwed this up, or if it's been like this a long time; usually in the VT I am logged in as root, because I'm usually fixing something or installing nvidia driver (as root) or whatever, so maybe I haven't done something *specific* lately that caused this, but I dunno.

What I've tried to test/figure this out:

1) copied <user>'s ~/.mc/ini file over to root to see if it caused trouble. It did not.
2) sorted & diffed `env` output of <user> vs root when in the VT -- basically the same for both.
3) temporarily removed all .bashrc, .bash_profile, .inputrc files from <user> -- still borked.
4) as root, tried sourcing the above files belonging to <user> -- no go, still worked fine for root.
5) verified the same $TERM variable is being used for both <user> & root, but tried other ones too - no useful result.

Not sure where to look next!

When inside X at my desktop, I can use MC as <user> or root, no problem. This is only happening in the VT.

Open to ideas

Sasha

I have discovered something "interesting" (for lack of a better word). The problem doesn't appear to have anything at all to do with anything normal-- it isn't a config file, or an environment variable, or an MC setting.

When I am running X (i.e. I've run `startx` as <user> and am logged into my desktop as <user>) the problem is apparent! When X is not running, the problem is not happening.

So, with my desktop running, I switch to an empty VT, logging in as <user>, and start MC, it's all screwed up, with Fn keys not working, direction keys not working, etc. Now, if I switch back to VT7 where X is running, and log out (killing X) and then go back to the VT where I am running MC, it works perfectly!

What to take away from this? Either my WM, or X itself, is for some reason interfering with keystrokes in the VT I guess.

Must talk to the WM devels and see if they have any insight.

Sasha

It's not the WM, as the same thing happens if I start TWM. Must be X??

Maybe someone else can try this and see?

-- Start your X like usual (I use `startx` from init 3)
-- CTRL-ALT-Fn over to an empty VT, and log in as same user your X session is running as.
-- start up `mc` and see if your arrow keys and function keys work.

For me, function keys do nothing, and arrow keys appear to be producing some sort of bash history sort of action, in the mc command prompt area.


ALSO, for the record: I'm still running Xorg-1.6.3 since I haven't yet got 1.7.5 to work.

I'm not an mc user but it seems to work.

The Xserver has been interfering with my keyboard on VCs for a couple of releases. I don't quite get the same symptoms as you do but I do get weirdness.

Firstly, I'm using a oem-rebranded chicony wireless keyboard and mouse that has a wireless usb receiver, which may be significant a I've not seen this happen on a machine with a ps/2 wired keyboard.

What I find happens is that if I startx and then exit back to the VC the system seems to miss keypresses. If I type very deliberately and slowly then the keypresses will be registered but if I type at any rate faster than one character every second or so then it starts to lose keypresses. It seems like it might be some sort of timing/syncronisation/timeout problem.

The really interesting thing is that if I restart X on vc7 then the other vc's will start working normally again. Initially I thought this was something to do with HAL which was why I tried adding the xorg.conf lines to disable hot-plugging, but I haven't tried completely compiling hot-plugging out yet so that may be an avenue to pursue at some point. If I just use the vc's and don't start X then the problem never occurs.

What's most interesting is that I get exactly the same problem when doing the same thing using OpenBSD 4.5/4.6 on this hardware, which suggests it's Xorg rather than Linux that's causing/triggering the problem somehow.

Whether this is in anyway relevant to your issue I'm not sure, but it sounds like it's kind of in the same area.

GazL, thanks for that feedback.

For comparison: In my case, I'm using a PS2 keyboard & mouse. I also have input hotplugging disabled in xorg.conf, and have not yet built an Xorg-server without HAL.**

If I have no X running, and I use mc in a VT, it all works fine-- but then, if I go to another VT and start X, then switch back to my still-running-and-previously-working-fine instance of mc, it now is borked. So it seems definitely X related.

I agree that it doesn't seem to be a "generally Linux" problem -- I have been using mc in the VT's ever since I began using Linux, so several years now anyway, and I've never found this problem manifesting. Trouble is, I really don't know *exactly* when it may have started happening-- i.e. was it last week, last month, or when I started using -current, or maybe the problem was there in -stable also-- I just don't know.

** I have reason to believe (read: "I've heard") that upcoming Xorg versions (from X.org) will again be HAL-free, but I don't know if it's true, or if/when it might actually happen- so nobody get excited..