okay, checking out the repo now.

oh, deleting old libs will of course break any ports you have installed from 13.x, so you might want to hold off on that

deleting old libs? if i understand correctly, make installworld would be the first thing that might do that.

installworld doesn't delete anything (though it will replace files)

there's a separate target for deleting files that existed in older versions but not the current one.

yeah, to my understanding what would delete old libs would be pkg upgrade… and after that's run through nothing should be on the system (foregoing custom port installs, which i don't have) that tries accessing those old libs. so I'm not quite sure what it is you're warning me about.^^

no, I mean base system libs

ah, i understand.

there are targets for make delete-old* in /usr/src to delete obsolete base system files

delete-old-dirs, delete-old-files, delete-old-libs and delete-old to do them all

and these targets exist in stable/13, but not main?

no, all versions

ah, parsed that sentence wrong.

the point is that if you installed any ports or packages from 13.x, those will rely on base system libs from 13.x

so i'll have to do pkg upgrade (possibly with -f) to fix my userland, correct?

and i should probably execute the delete-old-* targets, but only after i finished the kernel+base update and pkg upgrade.

so the thing about -current is that it's a moving target for ports, without a stable ABI you don't know if packages built a while back will still run

alpha.pkgbase.live is back! — (IPv4 only for now, need to figure out what vultr does with IPv6)

Title: Unofficial FreeBSD pkgbase repository

RhodiumToad: so, essentially, updating to 14 is only really an option if i build all packages on the same machine? :/

not necessarily, just that you can't expect things to "just work" the way that released versions usually do

pretty sure the handbook mentions this kind of thing

mhhh… i can't use the poudriere on my 13.1 homeserver to build packages for 14, can i? ^^

yeah, you can

er

no, you probably can't

huh, i always thought i can build for older, but not newer versions of freebsd…

okay^^

and the release schedule for 14 looks like it's already ~2 months delayed. so i essentially either upgrade at least also my homeserver so i have compatible packages or wait until release so i can upgrade all my infra at once without the unstable hassle…

phryk: OpenSSL 3.0 has been a challenge

it's the openssl update

also, to a certain degree, the LLVM update, which whacked a lot of ports

which llvm is in 14 now?

ah ye, i recently got a mail about that. glad i already passed sequoia maintainership on to vishwin, who recently finished updating it to 1.16 ^^;

meena0: i imagine openssl 3.0 will also throw a wrench into my plans to migrate to libressl at some point… :/

phryk: libressl and openssl 3.0 are probably a lot closer than openssl 1.1 and libressl

i hope it at least improves things instead of making them even more convoluted and opaque. :F

libressl has removed a lot of gunk years ago, that OpenSSL is only now deprecating

ah.

good riddance

yeah, that already seemed like an overdue removal of attack surface back when libreoffice threw it out.^^

err libressl :D

it's not just attack surface, it's also maintenance burden, and absolutely nonsensical for developers

but we have to support the mk68 architecture!1!!

why would I have to implement a TLS state machine in my program? why can't my library abstract some of that away??

I think they fixed it, but there was one version where if you tried to write data and it only did a partial write, then the next call had to pass the _exact same_ buffer (including the supposedly already written bytes)

which might be the most insane thing I've ever seen a library API do

is that pretty close to how the normal c api for writing to files works, too?

no

if you write() and it writes partial data, the app is free to throw away the data that was written and reorganize its buffer

(e.g. you might memmove() the unwritten part of the buffer down to the start)

but you still have to handle the case of a partial write.

or move it to a whole new buffer and use that on the next call

right, but openssl required you to not move the data in memory

oh. that sounds like it'd lead to horrible frustrations when debugging.

like you wouldn't believe

okay, i opted to forego 14 for now, even tho that means it'll take a few more months before i can switch boards between desktop and server again. still got enough to do on the server itself to pass the time, anyhow.

phryk: you could also try PkgBase for upgrading, but i wouldn't try that with CURRENT, without BEs, and you don't have ZFS anymore

meena: nah, i think i'll rather fix up my server stuff first. maybe in a month, i'll be antsy enough for this^^

i don't think i'll ever get over how route allows me to just kill all routes but has no way whatsoever to just show them so it's by definition something you have to do on blind faith… :F

well, i mean flush shows the routes it deletes, but that seems a bit late…

netstat -r

(or -rn, which is usually more useful)

true, but it still seems like a missing feature.

just found out... regarding browser-fingerprinting FreeBSD users are screwed...

User-Agent similarity <0.01%-<0.03% similarity to other users :D

ah

I've had a few websites give me error 403 when Linux and Windows users weren't getting such errors.

How odd, for a site to ban FreeBSD users. Probably unintentional.

well, I've had some weird issues with citrix web-client on FreeBSD firefox - was not a supported platform

but FreeBSD on chromium works...

crap-ware

Chromium on FreeBSD, no?

yeah :D

I would not be surprised if it was the former

xtile: Quite possibly intentional, tbh

But that's just a conspiracy theory. Don't take it for fact

it's fact for some sites, it's not about FreeBSD, rather not seeing Windows, macos, Linux in user-agent string

That's my guess, yeah.

Sorry, ignore that

… I mean, ignore the "Via" message that I sent (if you see it) before saying sorry.

grahamperrin: ferrovia?

(it is not seen)

hi there, quite often i'm finding regex in scripts like ${line%%=*}, but never found a good documentation for this

it is not regex

it is shell glob

can anyone give me a hint where can I foud detailed description of that csh/sh patters and siles

that one greedily cuts off text at the end of ${line} matching =*

`man 1 sh`

"sh(1)" manual page

parv, thanks

finally !!!!

THANKS !

nerozero: pubs.opengroup.org/onlinepubs/96999…ilities/V3_chap02.html#tag_18_06_02

Title: Shell Command Language

"Remove Largest Suffix Pattern."

xtile, excellent !

:D

thank you !

LxGHTNxNG, thank you :)

But yes, the man page also has the same info.

And if you mean you want information on glob patterns... in the man page they're at section "Shell Patterns"

yep

that is exactly what i need

and in the standard they're here pubs.opengroup.org/onlinepubs/96999…/utilities/V3_chap02.html#tag_18_13

Title: Shell Command Language

I am not to be thanked, nerozero

why ?

don't question it.

hi im using ungoogled-chromium and every 5-10 reboots it deletes all my preferences. its annoying having to set it up again and again. how can this be fixed

So i posted my PR 24hrs ago @ bugs.freebsd.org/bugzilla/show_bug.cgi?id=271938 , now i am making a thread in the forums too, for more help on this, i also plan to discuss this in the mailing lists. Is that encouraged or frowned upon?

Title: 271938 – Excessive heating on Lenovo Thinkpad E450 even with CPU @ 100% idle

Hi, I'm trying to start a Bhyve vm (using vm-bhyve) from a remote ssh connection and it starts and immediately stops with error 4. If I start the vm from the physical host it runs correctly. Is this a bug?

Hi, I'm trying to start a Bhyve vm (using vm-bhyve) from a remote ssh connection and it starts and immediately stops with error 4. If I start the vm from the physical host it runs correctly. Is this a bug?

RhodiumToad: Hey, i posted in the forum too, JFYI :) - forums.freebsd.org/threads/excessiv…d-e450-even-with-cpu-100-idle.89358

Should i share this issue in the mailing list too? or would it be considered too much?

Is ZFS protect against bitrot if you use a disk which is powered off after the backup was made and not running?

sozuba: if you post in more than one place, have cross-references

Agreed.

grahamperrin: got it. I have cross referenced my PR in the forum post, will do a reference of both on the mailing list when i do. Should i also cross reference my forum post on the PR?

Link from the bug report to both (a) the forum post; and (b) the archive of the email.

grahamperrin: okay, already done. Thanks :)

I'm in Bugzilla at the moment, I'll link to the forum whilst I'm here

grahamperrin: the bug PR has been already linked to the forum post.

or have i missunderstood what you meant?

annd thanks, i noticed you posted the forum refence in the PR

NyanBSD

|F9|: I'd like to subscribe to your Mewsletter.

MyQuiL: sure go to nyaaa.nyaa/mewsletter

;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 14804

You can join my chan, and read about mewsletter how I have eaten lol

I am joking, of course!

i have a small server with an html vnc console, and when trying to type my geli password during the loader, it treats enter key, delete key, etc as just another character. but any other time (even in the first bootloader), this does not happen.

UEFI or legacy boot?

hmm...i should know that but i'm not even sure.. it has gpart bootcode -p in the partition that it boots from, then loads /boot from a UFS partition

gptboot partcode

probably legacy

is there an ESP? (partition of type "efi") ?

no nothing like that

first partition is of type freebsd-boot

legacy boot then

OK

i haven't had this problem on other hosts, so i would think it's some encoding issue or some thing with the VNC

(by host i mean service provider)

you get the prompt "GELI Passphrase for %s " ?

since it's the -b option (main loader), it says "Enter passphrase for %s"

this is from loader.conf

it's also true that the virtual USB bus/keyboard appear to be loaded just after it asks

along with some other things. which is not what happens with other providers (the 'Enter passphrase for %s' is usually the final line, i think)

uh

can you show a paste or screenshot of the relevant lines?

let me see

this is on 12.4

with the geli -d option ('*' appear in the password prompt), i can see that the enter key, delete key, etc are added as characters in the password, and i can't have it press enter

that's in the kernel, not in loader

OK sorry for any confusion. i just meant this is the prompt that is configured in loader.conf

I don't think it actually is; it's just that it happens to be the same as the one hardcoded in the kernel

OK

it's using cngets for input same as any other kernel console input

I'd be tempted to try kern.cam.boot_delay=2000 in loader.conf to let stuff settle down before trying to access the disks

cool, i will try that thanks

hmm, same output/order of lines, maybe a longer delay

hm, odd.

possibly that delay isn't in the right place to affect this

you could try increasing it more (e.g. 10000) but if that doesn't help, might as well take it out

OK

hi

yeah seems that maybe this parameter does not have an effect at this stage

looks like it only delays the attempt to mount root, not things like geom tasting

hi guys!

i installed freebsd, gpu drivers, i3, everything works great

what do i do with it?

what do you want to do with it?

i dont know, can i game on it?

like steam or lutris?

how do i get bluetooth audio?

bluetooth audio is reputed to be possible with a bit of work, using ng_bluetooth and a virtual_oss driver

oops

Anyone uses FreeBSD as main server OS?

oh I wish

We use it for host OS on our infra.

I wish I could use it for everything.

Can I install it to macbook?

I am working towards making it my everything OS, hopefully soon

Nice

I just use macOS, recently got another mac

lol

I'm more wondering, if I can install FreeBSD to powerbook? That would make it at least a little useful.

I would maybe run BBS on it.

Also, is there anyway to see amdgpu temperature? forums.freebsd.org/threads/amdgpu-i…mperatures-and-or-power-usage.76245 has good discussion, but nothing concrete.

Title: [amdgpu] Is there a way to read AMD GPU's temperatures and/or power usage ? | The FreeBSD Forums

gnite all

answer my question someone someday pls

I shouldn't say "I am working" I amjust postung issues and geteing the hekp of the community to make it happen :D

posting*

F9 what was yoou rquestion?

"Anyone uses FreeBSD as main server OS? is that your question?

ah, just noticed.Good night

=D

To anyone who subscribes to the freebsd-current list: did you receive my email with the following title?

> Directory 1002/ missing from /var/run/user/

is there a way to tell netcat listen to shutdown if nothing is received within x seconds ?

all timeouts I see are more for sending and man page says they won't work in listen mode

grahamperrin: is it not showing up on mlmmj?

last1, See "nc -N" in the man page. (OpenBSD also has a -q SECONDS option.)

last1, Also see the "man 1 timeout" command.

sozuba: been using FreeBSD as my server OS since around the time VIMAGE was about to be merged into GENERIC

meena: VIMAGE GENERIC are words that speak Klingon to me :)

but i got the gist

VIMAGE is kind of misnamed, IMO

the ony VIMAGE I can recall is for an android ROM patch, ahaha

RhodiumToad: or rather, the actual thing is way less ambitious than the initial idea

ah

(but that's with all cores pegged at 100% on a hot day)

good cooling?

well, seems acceptably good

Well, VIMAGE changed scope; the original idea was much broader.

As far as modern nomenclature, calling it vnet is correct.

community.amd.com/t5/drivers-softwa…reebsd-and-amdgpu-temps/td-p/475396 -> no response, Is amd always like this?

Title: FreeBSD and AMDGPU Temps - AMD Community

could someone confirm if wiki.freebsd.org/TuningPowerConsumption#A2._Screen_.2F_Video are still relevant to 13.2-RELEASE?

Title: TuningPowerConsumption - FreeBSD Wiki

Title: Solved - Intel i915 rc6 power saving enabled? | The FreeBSD Forums

got to go ppl, bye, have a great time

hmm, i've found that for some reason, when one removes a passphrase on a geli boot volume, a passphrse is still requested during boot. it appears that reinitializing the geli volume is required

you removed it how?

i forget the exact line, but something like: geli setkey -K $KEY -P

then i confirmed it was ok by attaching with -pk

it then notes that the old metadata may still be present in the backup location, i think

how are you booting it?

i set gpart configure -b on the partition containing the geli

without that, it will try mountroot without attaching, i think

but now when i reinitialized it with -PK it does not request the passphrase

Hey, I have a quick question. FreeBSD's NFS is implemented in userspace, yes?

no

thanks

In TrueNAS, which is based on FreeBSD 13.1, I can create jails of FreeBSD 13.1 and 13.2. How come I can create 13.2 jails? I was told it's only possible to create jails of the same system version or lower.

Possible. Unsupported

Thanks, so I'll make 13.1 jails then.

anything is possible, how well it works is another story

otoh, 13.1 will be EoL soon. Is TrueNAS gonna update?