-
mason
Not Tested was the traditional one.
-
rtprio
should i start moving my systems to pkgbase
-
mason
khy,.khg~k;il
-
mason
('
-
mason
Not Tested was the traditional one.)\\
-
mason
)
-
mason
With their two examples I'm struggling to see what's matching.\\\\\\\\\\\\
-
mason
'
-
mason
)~~~~~
-
mason
---9
-
mason
0
-
mason
very sorry\
-
rtprio
is there something wrong with your 'enter' key, son?
-
mason
Mm. Was swapping keycaps and my mouse drifted over my IRC window.
-
rwp
mason, Usually people blame it on their cat rather than on their mice but you do you! :-)
-
mason
My cat *was* on my desk, as it turns out.
-
CrtxReavr
If my dog ever discovers the keyboard, you will know it.
-
ketas
or children
-
pertho
is WINE still supported in FreeBSD? wine-devel 10.10_1 seems to still do the pkg32.sh song and dance.. but WINE 10 should support both 64-bit and 32-bit WINE prefixes
-
hodapp
wow, didn't realize it was ever supported
-
pertho
are there any special sysctls I need to set to make WINE stuff work?
-
pertho
WHen I try to run the WINE app (works in Linux+WINE) I get: wineserver: file_set_error() can't map error: Cannot allocate memory.
-
pertho
and yeah I turned off aslr in sysctl for all the little knobs and dials referring to wine
-
pertho
kern.elf32.aslr.stack, kern.elf32.nxstack, kern.elf64.aslr.stack, and kern.elf64.nxstack .. all set to 0.. tried increasing kern.ipc.shmall... nothing changes
-
jbo
pertho, let me know once you figure it out.
-
jbo
I can't even get winetricks to run
-
pertho
wow
-
pertho
-
pertho
but.. none of the things in it help
-
pertho
jbo: you on 14.3 or 15?
-
jbo
stable/14
-
jbo
(so 14.3-ish)
-
CrtxReavr
Is it time for 14.x?
-
CrtxReavr
I've been a 13.x luddite.
-
CrtxReavr
Looks like 13.5 is still supported.
-
pertho
0024:err:winediag:WSASocketW Failed to create a socket of type SOCK_RAW, this requires special permissions.
-
pertho
can FreeBSD allow normal users to create raw sockets?
-
pertho
all I see in sysctl is for jails: security.jail.param.allow.raw_sockets and security.jail.allow_raw_sockets
-
beastie
pertho: make the program setuid root.
-
ivy
making wine setuid root would not be a great idea
-
hodapp
but then it's just like old-timey Windows
-
hodapp
didn't you want compatibility?
-
beastie
windows alone runs setuid :)
-
pertho
which program?
-
pertho
wine is just a shell script
-
agent314
is 14.3-RELEASE already ready for freebsd-update from 14.2?
-
agent314
it's saying the following components dont seem to be installed: world/base-dbg world/lib32-dbg
-
rtprio
so you don't have debug base or lib32. no big deal
-
agent314
idk seems like it's fetching and applying stuff
-
agent314
i think 14.3R was supposed to have come out in june
-
pertho
yeah it came out like a month ago
-
agent314
oh yeah it's merging a new sshd_config i think we're good
-
Macer
Is it merging new sudo? ;)
-
rtprio
sudo is a package not in base
-
rtprio
so why would it do that
-
pertho
pkg install doas :)
-
pertho
or just use su like a crusty UNIX greybeard :)
-
ZedHedTed
su is what the cool kids use xD
-
rwp
Just remember that when using su always use it "su -" with the option to load root's environment which includes PATH.
-
deimosBSD
well, zfs snapshots just saved me. accidentally wiped a filesystem--which happened far quicker than I expected.
-
rwp
deimosBSD, ZFS FTW! Having a safety net is the only way to go.
-
oxbar
Do pr's get pushed aside for more important issues or how does it work ?
-
ivy
oxbar: there is no queue of PRs to push anything aside from, developers work on PRs as and when they feel like it (or if they're being paid to)
-
oxbar
nice
-
ivy
the foundation does prioritise how to spend its money, which might affect which PRs get fixed, but they have no control over freebsd itself so their only influence is if they pay someone to fix something it's more likely to get fixed. like with the current wireless changes
-
oxbar
Got it
-
oxbar
you a dev ivy ?
-
ivy
i am a src committer, yes
-
oxbar
Question.. why does sound work and a lot of other things work out the box with older hardware.. some of them have almost the same chipset example for sound but newer laptops/desktops don't work and you got to fiddle with it ? is it because freebsd is trying to catch up with the latest ? how would you know what laptop/desktop to get if you want to
-
oxbar
try freebsd ?
-
oxbar
it will be to late to trake the desktop/laptop back if you know what i mean
-
ivy
oxbar: i don't know much about sound, but you might ask christos@ as he's been doing a lot of work on that recently and might be interested in reports of things that don't work
-
ivy
but HDA sound is quite complicated since it only defines how the generic interface works, vendors are free to wire things together in all kinds of unusual ways which makes it harder to support everything
-
ivy
even the same chip could be wired different in different laptop models
-
ivy
as for what laptop to buy, i don't own a laptop so i don't really know much about that either but i'd probably look for success reports from other users, or perhaps ask questions@
-
ivy
the foundation is paying for work to improve Framework laptops in particular, so perhaps those might be a good choice when that's done
-
oxbar
Thanks for the explination.. I think end users want to try freebsd but when there hardware dosen't work they jump ship. i always thought in freebsd if it worked it worked if not file a pr and just wait.. i have 5 laptops right now and like to go between them for some reason lol .. one of them i really want to work i filed a pr on and i'm just
-
oxbar
waiting. using my mac in the mean time. Thanks again
-
rtprio
waiting for what?
-
rtprio
for pr's to get fixed?
-
oxbar
my bad i was AFK had to go rake the yard and wash dishes 0_o
-
oxbar
rtprio: waiting for the pr to get fixed.. its not that big of a deal.. once its fixed i will reinstall to freebsd and test it out
-
oxbar
-
oxbar
The forums are amazing!
-
CrtxReavr
Well, they're okay.
-
ivy
i think i have a forums account but i've never used it... they ban all discussion of -CURRENT so there's not much for me to talk about :-)
-
CrtxReavr
Honestly, if you need help, there's no beating IRC.
-
TommyC
The forums ban discussions of -CURRENT?
-
ivy
TommyC: their logic is such discussions should take place on the mailing lists where the developers are, which is understandable
-
CrtxReavr
I know there's a Discord "server," which I've joined, but not paid much attention too yet.
-
polarian
ok freebsd networking question, vm and tap, all guides are bridged, I am doing routed, however I have ran into something weird I cant wrap my head around. so I have the ethernet device within the vm which has 192.168.254.4/29 with gateway 192.168.254.3/29, the tap interface has address 192.168.254.3/29, this is a jail and has the default gateway of 192.168.254.1/29, the jail can access the
-
polarian
host/internet just fine. However if the vm tries to ping 192.168.254.3/29 it doesn't get a reply, however when using tcpdump on tap0 I can see the echo requests...
-
polarian
pf has a pass all rule, thats it, no blocking
-
polarian
pass all flags S/SA keep state only
-
polarian
the vm has no firewall either... and again the packets can be seen coming in on tap0
-
ivy
polarian: you should assign IP addresses to the bridge interface which contains the tap interface, not the tap interface
-
polarian
ivy: I have no bridge
-
ivy
oh, i read that as "all guests are bridged"
-
polarian
that would be the easy route :P
-
ivy
polarian: in the jail, show the output of 'ifconfig <tap interface>' and netstat -rn, and the output of tcpdump
-
polarian
ah thought you would ask for this, alright one sec let me pastebin both
-
polarian
-
ivy
polarian: your routing table looks odd. why is 192.168.254.0/29 routed via jail0 if 192.168.254.3 is assigned to the tap interface?
-
ivy
if you're tried to assign addresses in the same subnet to multiple interfaces, that is not supported
-
polarian
ivy: would explain the issue then, to explain it jail0 is a epair, one of the pairs is on the host with .1, one in the jail with .2 this works fine, tap is .3 and the if within the vm is .4
-
polarian
I did discuss this with some Linux networking folks about a month ago, and none of them said it was abnormal soooooo
-
ivy
polarian: right, that won't work. you need one subnet to communicate with the host then a separate subnet to communicate between the jail and the VM.
-
polarian
ughhh
-
polarian
I was told *NOT* to do that though
-
polarian
as then you are nating twice...
-
ivy
i can't speak to why someone told you that but it sounds very wrong to me, no IP router should accept this configuration as it doesn't make sense
-
ivy
no, you don't need to NAT at all
-
polarian
oh wait true
-
ivy
except if you have to NAT the packets headed to the Internet, but that's still only one NAT
-
polarian
anyways easy fix, split into two /30 (or /31, but not sure if point to point will work in this use case)
-
ivy
fwiw, if you don't want to allocate a /30 or /31 to every VM, you can use /32s instead
-
polarian
I dont see how that will work
-
ivy
but it's a bit harder to set up so probably not worth it unless you're short on addresses, or you already use IPv6 (in which case you can route the /32 via the IPv6 address)
-
polarian
I am using RFC 1918 ranges, and holding all IPv4 on the router
-
polarian
most efficient way to allocate the IPv4 addresses...
-
polarian
means NAT... but also means I can share addresses
-
polarian
yes I can and do make two servers which have no conflicting ports share the same address
-
polarian
with IPv6 its not an issue, but with IPv4
-
polarian
its efficient
-
polarian
IP wise...
-
ivy
/31 works fine in freebsd, btw, including on ethernet. i think linux only got support for it surprisingly recently
-
polarian
ok cool
-
polarian
not like there is a limit of RFC 1918 addresses, buy why waste addresses when you dont need to :D
-
ivy
remember you also need to route the new /31 to the jail host, probably easiest to just allocate it a /24 and route that so you're not having a bunch of tiny routes...
-
polarian
right so now I understand the basics of networking I can go fix it :P
-
polarian
oh wait that was my stupid mistake... and why I need to plan shit
-
polarian
I have already done that, I just didn't subnet the block routed on the host
-
polarian
within the jail
-
polarian
ugh this is why you dont leave projects have finished and come back 1 month later!
-
polarian
thanks ivy for fixing my stupidity :D
-
thresh
Hi, I'm getting the following while trying to build a perl-dependent port on a freebsd 14.3 vm, using quarterly (2025Q3) ports: ===> p5-Error-0.17030 Invalid perl5 version 5.36.
-
thresh
`pkg install perl5` says I have the most recent version.
-
thresh
I'm also using quarterly pkg collection (which is a default anyway, I'm on AWS VM). Doesnt look like there is an updated perl5 version. Is it supposed to be like that?
-
ivy
thresh: 2025Q3 doesn't have Perl 5.36, but i don't see anything in lang/p5-Error that would depend on it. is it possible you have set DEFAULT_VERSIONS=perl=5.36 in /etc/make.conf?
-
ivy
(if so, you need to choose 5.38, 5.40 or 5.42 instead)
-
thresh
thanks ivy, I don't have anything in /etc/make.conf (it does not exist)
-
thresh
I do have perl5 package installed via pkg repos; and it's indeed 5.36
-
ivy
thresh: what does 'grep PERL5_DEFAULT Mk/bsd.default-versions.mk' say?
-
thresh
PERL5_DEFAULT?=5.40 (plus some extra stuff)
-
ivy
and what version of the perl5 package do you have installed?
-
thresh
I do have perl5-5.36.3_3
-
ivy
what package repository are you using? from /etc/pkg/FreeBSD.conf
-
thresh
it's the quarterly, url: "pkg+https://pkg.FreeBSD.org/${ABI}/quarterly",
-
ivy
it seems like the quarterly builder hasn't built 2025Q3 yet, so you should either use 2025Q2 (i assume) or else build everything from ports
-
thresh
is there an url I can track the build process for 2025Q3?
-
CrtxReavr
I would argue perl is outdated period. . ..
-
CrtxReavr
Who remembers when perl was part of FreeBSD's base?
-
thresh
I'm building a piece of software (proprietary), and would like my customers to be able to use the binary I provide. Wouldnt be really possible if I'd mix and match stuff
-
ivy
shipping proprietary software that depends on ports is probably not going to work very well. although if it's just perl, you could build the package yourself and just depend on "perl5" rather than a specific version
-
thresh
Indeed, it's a bit of a pain. I can see at least a shlibs dependency to libperl.so.5.40, and "dep" in pkg MANIFEST to a specific origin/version (no idea if it's enforced by pkg)
-
thresh
(I did install perl5 from latest pkg repository to test things out)
-
thresh
so my best bet is to wait until things settle down; hence my question on tracking the status
-
ivy
if you have binaries that link to libperl, you're going to find your software breaks every time users upgrade to a new quarterly branch, or just change perl versions. i'd suggest shipping perl with your software instead... it's not an ideal solution but it at least works
-
thresh
yeah, that's exactly what we experience (and we depend on more stuff from ports)
-
thresh
so far it's been easier to just rebuild our stuff and push out updated packages than to vendor our dependencies
-
thresh
I'm looking at
pkg-status.freebsd.org/beefy20/buil…2amd64-quarterly&build=c4772dbffe33 and it seems perl5 was in a queue - I can see it in "queued ports", but it's not in any other section, and the Build Status is done