Oleg: sysctl kern.evdev.rcpt_mask might be what you're looking for

Oleg: try setting that to 12, which should direct keyboard and mouse events to the hardware devices rather than to kbdmux

Oleg: (programs not using evdev / libinput should be unaffected)

RhodiumToad: oh, thanks for responding! Let me try it... And as I already mentioned, the mouse event was directed to the dell usb mouse itself..

I think if moused is not running, there's nowhere else for mouse events to go

actually, moused is running: /usr/sbin/moused -p /dev/ums0 -t auto -I /var/run/moused.ums0.pid

what value did you have rcpt_mask set to?

You just told me to set it to 12, but earlier it was 6

(it's common these days to need to set it to either 6 or 12 to get X to work)

right, 6 directs mouse input to come from the mouse hardware rather than sysmouse,

but leaves kbdmux enabled, while 12 has both keyboard and mouse input come directly from the hardware device

the reason the mask is needed is because otherwise programs using libinput might get events twice

it's possible your app is ignoring kbdmux input as an (incorrect) bugfix for duplicated input

RhodiumToad: Thanks! That fixed that keyboard issue I had while running kodi on the bare console. I don't know why this problem didn't manifest itself when I was running kodi in the wayland compositor environment.

RhodiumToad: I am such a n00b. As I already mentioned, earlier I did silly things such as adding hint.kbdmux.0.disabled="1" to device.hints or removing the atkbd device from the GENERAL cobnfig.

RhodiumToad: with the kbdmux input, if I pressed one key immediately after pressing another one (which is what normal people do), "events" (as if keys kept getting pressed) kept happening by themselves in kodi.

but it works ok with the new rcpt_mask setting?

RhodiumToad: Yes! Thanks!

RhodiumToad: and on the bare console, I had to edit devfs.conf to change permissions for event devices because kodi was complaining. (Before I did that, neither the mouse nor the keyboard worked in kodi). But in the wayland enviroment, kodi didn't care about permissions.

I find it a bit puzzling because both the wayland compositor (sway) and the bare console want access to event devices.

I mean, if an app relies on libinput

brb

I was just going to say but whatever he left, that wayland doesn't have an indirect layer to events like X11 does. Since Wayland is designed to be "minimal" as a display protocol, it needs to access input devices for the input events. A feature, not a bug.

and as a result, if you want to manipulate events for configuration purposes, it needs to be done via root.

urrrh, (ungoogled-)chromium's hardware acceleration can really bog down KDE's compositing. Would there any reasons to keep hardware accel enabled in general?

just tuxracer

Not sure if rtprio was just being funny or incredibly rude/disingenuous

kenrap: i think… funny… but that as 8:28 my time, and i have no judgementability at that time.

I hope so too

pkgbase report! freebsd/freebsd-doc #219

meena: and good morning! o/

Title: add a 2023q2 report for pkgbase.live by igalic · Pull Request #219 · freebsd/freebsd-doc · GitHub

2 reports in 3 days, and just a week or two later than promised.

now to get back to work???

meena: just a suggestion in your pkgbase guide, for the `kld_list="filemon"` metamode part, would it be better to mention `sysrc kld_list+="filemon"` as a copy/paste action for the reader?

*filemon

kenrap: if you have a codeberg account you can submit a patch so it doesn't look like i'm the only one doing stuff ;)

Sure thing. I signed up for one recently too.

awesome thank you!

👍

meena: boom, PR is up!

Also, don't mind when I prefix my branches with my username. I do it as a scheme to help me with tab completing my branches and to differentiate from remote branches that I checkout

merged and deployed

Thank you!

The latest update of nrpe broke something, nrpe is starting properly but according to the page status_services.php, it's not started

Everything works properly except that pfSense see the service as not started even if it's started

wrong channel, lol, my bad

good morning

Do people still use tcsh here?

wizardsmoke: i use it as my default shell for root. for my own User i install fish

(can't stand sh, but i think that's not FreeBSD's fault. more likely it's Solaris)

I am having some trouble with a script. I need to source it before (or when) I drop into my preferred shell to set some environment variables but its complaining.

Let me paste something for you so you can see the complaint.

Title: View paste BUMOS

so that first test is what complains. I call it in .tcshrc with `source /path/to/script.tcsh` and I've printed sourced_script to the terminal. It appears to be blank the whole time.

If I manually source it in the shell it works, but not when I source it from the rc file

interesting

Yeah, its real bizarre. So, obviously I could source it in tcsh and then open my shell but that's tedious to do every time I open a shell

Plus, I'm sure this person could have just implemented this script in sh which seems to be the most universally compatible.

hm, maybe if I just copy and paste their script into my tcshrc it would work

wizardsmoke: I keep wanting to use tcsh but it keeps besting me.

if i have n jails, each with packages installed, what's the best way to update them? run pkg upgrade n times, once per jail? or is there a way to fetch the repo catalogue once?

running it n times is what i'm doing, so just checking that this isn't frowned upon

markmcb: if you have categories of jails, i would run one canary per category, and automate the rest

is it spam on developers@ 50 mins ago?

sure it is

occationally that does happen, yeah

thanks

postmaster@ does a pretty good job of keeping on top of it

if I have ifconfig_mce0_ipv6="inet6 accept_rtadv" do I still need ipv6_enable=”YES” in my /etc/rc.conf

crb: Nope.

w2/32

hmm ok then I'm confused I have a freebsd box that can ping my mac over ipv6 put not the other way around

crb: ipv6_enable for /etc/rc.conf is deprecated.

crb: Have a firewall running?

yes but not between these two systems, only to the outside world

crb: Could it be a gateway issue?

these two systems are on the same network

I wouldn't think they'd be using a gateway

Are you specifying the interface to use for ping?

no but I am specifying he IP address

crb: Can you try: ping6 2003:b8c::01%int0 ?

on which machine

Of course, changing the v6 address and %int0

The machine you're trying to ping from.

It'd be the remote v6 address and the local interface that has the v6 address you're trying to ping from.

ah that makes sense I was using ping on the mac with and IPv6 address ping on the mac doesn't accept -6 IPv6 and I didn't know there was a ping

You can also try the v6-local addresses and see if they work.

ping6

ek: now it works, thank you

crb: Great to hear!

i just paid $30 for a 2 year domain name registration. did I get screwed?

Just looking at the registration page at dnsimple.com/tlds it seems that it can vary greatly depending on the choice

so a .com is 14.50 while a .dealer is more than $2000 with three zeros (I assume those are the first year cost)

14.50 for a .com domain sounds rather expensive

s/sounds/is/g

still haven't found anything to do with the domain i registered back in 2021