the issue is that luajit and luajit-devel conflict on install, and some consumers use one and some use the other

oh, the gdal one is because gdal depends on postgresql-client of the selected postgresql version, which isn't necessarily the one in use

all of postgresql-* needs to be reworked to allow concurrent installs a la Debian, but that's not been done yet

i see

(by "not been done yet" I mean it's been talked about for 5+ years without anything being done)

I wish the same was available for MySQL/MariaDB. The only trick here is not to set up network access. That's how MacPorts has implemented it.

RhodiumToad, I created patch for obs-studio - check bugs.freebsd.org/bugzilla/show_bug.cgi?id=271085

Title: 271085 – multimedia/obs-studio conflict with neovim due to luajit dependency

I found a strange thing today.  that zfs dedup is very IO intensive and generally not recommended for anything. It will make more writes/reads on disk than the original data to be wrttien in first place

why is that strange?

I thought it would use cpu .. ok but will not write more than the real file size...

its saving space at cost of writes, cpu

what's odd is it seems to happen before ~/.zshenv is processed

... what does?

sorry, i think i got disconnected when i posted my comment, i wrote:

i have a serial over linux connection to my freebsd server. when i login with my user, it logs in and hangs after motd. if i ctrl-c, i'm back at the login prompt. root seems to works, so i'm guessing it's something zsh related. i use the same zsh config on linux without issue. any common "gotchas" that might cause this? if not, i'll keep hunting.

markmcb, Add "set -x" at the very top of ".z{profile,shrc}" files to better observe

have you snooped on the serial line to see what's being sent/received?

not sure how to snoop. parv, i'll try that.

parv, where should i expect to see the effect of that? on the console?

markmcb, Wherever zsh runs; I suppose on Linux side

markmcb, Note that version 5.0 of zsh lacks "pipefail" option in case you have that set

emitting the motd is the responsibility of login(1). The next thing that it does after motd is to check for new mail, which is just an fstat of /var/mail/$USER

and the next thing after that is the exec of the shell.

you might try control-T instead of control-C, since that should show what process is running

sysctl kern.tty_info_kstacks=2 will also show the kernel stack on control-T which makes it easy to see what it's waiting on

ctrl-t shows: load: 0.28 cmd: zsh 5389 [ttydcd] 276.01r 0.00u 0.00s 0% 3508k

ttydcd suggests it's waiting for carrier detect on the serial line.

perhaps the serial line was not set to ignore modem controls?

ah, perhaps that's it ... i kinda breezed through the SOL setup when i first installed

i'll double check my settings

note that setting is on the freebsd side

What serial program are you running on the Linux side to connect to the FreeBSD serial terminal?

it's a Serial over LAN link, using "ipmitool sol activate" in Debian

and what serial line type was set in /etc/ttye on freebsd?

/etc/ttys

/etc/ttys: ttyu1 "/usr/libexec/getty std.115200" vt100 on secure

you might want 3wire.115200 rather than std.115200

how to i reload that config after changing it? or is it automatic on next connection?

*how do

init q

(as root)

bingo

that was it

thank you so much, not sure i would have solved that mystery alone :)

the 3wire entry just tells it to ignore the modem carrier-detect / data-set-ready signals, normally opening the serial port blocks until carrier is detected

ah, i see. what doc would have more info on that? would be good to add to my reading backlog.

getty, gettytab, termios(4)

i can't find a reference to 3wire in those docs or ttys(5), and apropos lists nothing

the gettytab manpage describes the format of /etc/gettytab

3wire is an entry in that file

I'm impressed you knew "init q" rescans the ttys file.

I have to keep detailed notes of everything I do or I forget by tomorrow. I'm always envious of those who can retain nuggets of knowledge like this.

init q is actually an old SysVism.

all it does is do kill -HUP 1 but in some respects it's a bit safer

that's cool. i appreciate your insights.

btw, where would ctrl-t be documented? ... is that a unix thing? don't recall that one in linux

seems really useful

that's a BSD thing. it's the STATUS control character in stty

when the terminal sees that character when in canonical mode, it outputs a line direct from the kernel showing the system load, foreground command, wait channel and cpu/mem usage

if terminal signals are not disabled, it also sends SIGINFO to the foreground process group, and many programs respond to that by printing their own status/progress message

oh, I should have linked this before: docs.freebsd.org/en/books/handbook/serialcomms

Title: Chapter 28. Serial Communications | FreeBSD Documentation Portal

thanks for all the help! it's getting late here so i think i'm going to call it a night.

i'm on 13.1, and i've got a src.conf that includes "WITHOUT_TOOLCHAIN=set". it allows me to build a minimal FreeBSD for a small server, and it builds quickly, because it doesn't bootstrap llvm, it just uses the base system's included compiler.

however, if i try to build 13.2 on the same system with the same src.conf, it will always insist on bootstrapping llvm. this is something i can't get around when trying to cross-build between releases, right?

(i guess i could live-boot into a 13.2 image and build on there. quite a hassle tho.)

trying to sign up to oracle cloud and getting "Your transaction was declined." was declined for my credit card… I loathe that i have to add a credit card for a "free account"

what i just love is that the website says: "Your transaction was declined." and i have to look into the console to get something more better, like: "We are unable to authorize payment. Please contact your issuing bank or provide another credit card or an alternative form of payment (Error Code 478)." (there is no alternative form of payment)

meena: I got that too yesterday, after it looked like it had gone through

i can't even find 478 as an error code Visa would issue

okay, it means Strong Customer Authentication required, which means 3D Secure, which they didn't ask for… funnnn! love debugging shitty websites in the early morning

danluu.com/everything-is-broken ← i should get back to protocolling this.

Title: One week of bugs

okay, their support chat is also broken lol

NEXT.

meena: but larry ellison is so rich! it could only be because his products are superior!

Demosthenex: i wonder if he's rich enough to float in the middle of the Pacific

actually, if i credit oracle with anything, it's taking customer money. i'm amazed they have a bug in payment ;]

Demosthenex: they do?

Demosthenex: Ireland is a tiny market, and no one cares about it

if they had a DC here i bet it would be different, but only Microsoft, Google Abs AWS have DCs here

and by "here", i mean in Dublin probably. because out here where i am, we don't even have IPv6

we (Siemens) have a data centre in Shannon

but you have so many megacorps in ireland! how could you not have dcs?

hello there i have a problem i run a system freebsd 13.2 and i try run inspircd inspircd-4.0.0a22 and i get error when i run it without a module geo_maxmind then everything is good but when i run with a module i get error please if someone can help to me bsd.to/kJBA

Title: dpaste/kJBA (Plain Text)

it looks like this is a bad drive

paulf: do you have a cloud i should support?

hi there, any idea how to use less with "colored" output text?

nerozero: -R

meena, BLESS U!

thank you so much for always quick replies !

RESPECT !

nerozero: man less should tell you (i had to verify my response was correct ;)

-G with linux

I was looking for -G and didn't found that .. sooo my bad

-G is hilight search on my Linux laptop

maybe -G implies -R on Linux?

possible

hrm… we do seem to have -g and -G

someone please?

meena: no we're only just starting to get into cloudy stuff, many customers are otally paranoid and will never use cloud

paulf: well, if you wanna implement your cloud-init datasource give me a shout

paulf: also: I'm from austria and worked in Germany for a some time, so, yeah, I know.

it's usually the Japanese that are the most paranoid

but many others are too

i didn't realize Siemens was … big in japan 🎵

Hi, got a FreeBSD VPS rented online, getting its own public IP through DHCP pool of the server provider. How to see the IP address of the DHCP server? How to know what IP gives my VPS its IP address?

tercaL: To what end?

just trying to *whitelist* the DHCP server's IP address in my PF.conf file.

on my VPS

I assume you're talking about DHCPv4.

jkc: Yes.

Right.

Then don't bother. Pf won't filter that.

how do i chflags a file to immutable?

There's zero examples in the man page

tercaL: With DHCPv4, IP addresses aren't involved until you have an address. So how can you filter by IP address before they become involved?

jkc: My problem is; when I type "dmesg -a", I see:

"Jul 4 04:59:09 myhost dhclient[18366]: send_packet: Permission denied" and too many messages afterwards: Jul 4 05:11:29 myhost syslogd: last message repeated 18 times

So?

When I disable PF, it disappears.

When I enable PF, it re-appears.

So?

jkc: I'd like to know what's my mistake in my PF configuration, and what blocks the dhclient through PF.

The mistake is that you're blocking outbound traffic that you shouldn't be. Stop trying to whitelist outbound traffic.

Administratively, it's a horrible practice.

I don't have any blacklist/whitelist parameter or table in PF.

jkc: When I enable this passing rule, the message disappears again: "pass log quick on $ext_if proto udp from any to any port = 67" So it seems PF somethow interfering with DHCP server in the network. No?

tercaL: Stop doing that. Just `pass out quick on $ext_if`.

You're probably missing the "allow outbound traffic" rule.

jkc: For "pass out" I have two rules; 1- For Time, DNS: pass out quick on $ext_if proto udp from ($ext_if) to any port { ntp, domain } AND 2- For SMTP and Rspamd: pass out quick on $ext_if proto tcp from ($ext_if) to any port { smtp, www, https }

Isn't this a correct approach in terms of security, to not pass out everything - all ports to anywhere?

With "pass out quick on $ext_if" I think, I'd remove those two rules, as your suggestion allows communication of any port towards outside.. No?

Personally? No. I avoid outbound filtering. I find no value in imposing a control on the system via the same system that could remove the control if it were compromised.

The local system is filtering the traffic. If the local system is compromised, the attacker would just disable the firewall. You're imposing a lot of functional problems for little benefit.

tercaL: you can see it if you have tcpdump running when starting dhclient, but I'm not sure what the point of knowing the address is since the important parts of DHCP are done via broadcast.

Outbound ACLs aren't even remotely best-practice, unless you're dealing with systems that involve national security and that kind of threat level.

Got it! Some really good suggestions and super cool explanations! Thanks a lot jkc and debdrup

meena: don't you mean schg ?

tercaL you could also run dhclient in the foreground to see what the dhcp server is in that case

i don't necessarily have a credible opinion on it but i personally like outbound ACLs with whitelists, assuming it's easy to specify the whitelist...helps prevent exfiltration of data if there were some non-superuser comprimise

and for intnet -> intnet packets, it should help isolate such a compromise

jmnbtslsQE: schg needs root. i just want to set that on one of my files

I could also disable saving known_hosts

someone please?

hello there i have a problem i run a system freebsd 13.2 and i try run inspircd inspircd-4.0.0a22 and i get error when i run it without a module geo_maxmind then everything is good but when i run with a module i get error please if someone can help to me bsd.to/kJBA

Title: dpaste/kJBA (Plain Text)

ASHER: did you already open a bug on that?

that's not running that's building

yes and no boday try to help

ASHER: i reassigned it to the correct person

ok

Title: bug inspircd version 4 in freebsd 13.2 · Issue #2044 · inspircd/inspircd · GitHub

yes i told to him but he dont want to understand

is have a problem with a module when i run it

ASHER: the bug is in upstream, or rather, a dependency there: maxmind/libmaxminddb #318

Title: Misuse of _POSIX_C_SOURCE feature macro · Issue #318 · maxmind/libmaxminddb · GitHub

318 – matcd driver not on boot or fixit floppy FDIV012 bugs.freebsd.org/bugzilla/show_bug.cgi?id=318

ASHER: Sadie sounds like a "she" to me

i dont what a problem with this a man or women whatever anyway when i try run the commands like it gmake CXXFLAGS=-I/usr/local/include LDFLAGS=-L/usr/local/lib install i get error

i dont know if this my system or this a modules

if is have something else commands to try run?

ASHER: the problem is that this module doesn't build (on FreeBSD 13.2?) because one of its dependencies does something out shouldn't do

without a module is run good

but with this is have problem like in C++ i dont know why

we could try to patch that dependency, because I don't see any commits since October last year: github.com/maxmind/libmaxminddb

Title: GitHub - maxmind/libmaxminddb: C library for the MaxMind DB file format

Title: FreshPorts -- net/libmaxminddb: Library for the MaxMind DB file format used for GeoIP2

i install that

I know, but you have to patch it

i need run it in folder lib?

github.com/maxmind/libmaxminddb/blob/main/include/maxminddb.h#L8-L18 these ten lines need to go

Title: libmaxminddb/include/maxminddb.h at main · maxmind/libmaxminddb · GitHub

what i need to do copy the code or run it beacuse i am not see nothing abour run it in freebsd

and instead, you need to add that to the CFLAGS of geo_maxmind

ASHER: do you speak any other language other than English?

ohh sorry my english is very bad on sec

I asked in which folder I run it because I don't see commands for the operating system I'm using

meena> and instead, you need to add that to the CFLAGS of geo_maxmind how i do it is have commands please for that?

nothing to run. you have to delete those teeth lines from the .h file in your system. (somewhere in /usr/local/include)

ok

what about the CFLAGS?

ASHER: I'm typing this stuff on mobile phone. you're sitting in front of a computer. if you don't know how to modify a build system and patch a file, need to learn a lot before you're allowed to run an IRC server.

ok i will try it thanks my friend :)

meena i try to do what you explain and is not working

ASHER: I'll try to submit a patch when i get back to the computer, or you could wait for sunpoet to wake up

let's give them another four hours of sleep

but is not have there nothing configure :S you give to me the patch thanks but from here what i need to do you are good in computer me not my friend and what i am try to find this only good expalin

I have create the boot, the /, the swap and the /usr/home partitions by hand using Slackwares64's fdisk, however, in the step where you set the partitions to install the new system, the bsdinstall shows the error: "Error mounting partition /mnt. mount: /dev/ada0p3: invalid fstype: invalid argument"

I have tried with other installation images like the memstick, dvd, netinstall, etc but I still have the same error message

Why?

This is a 2TB blank dis HDD

The partitions were created using the GPT partition scheme

hello

Maybe this is a bug of the bsdinstaller program

I need some help I installed freebsd on my computer

another one downstairs and it says host not found

:'(

I can't pkg update or anything

asarch: Um. So, FreeBSD won't probably be happy installing on the filesystems a Slackware environment can create.

Oh

asarch to make us happy you need to convert 100% not 70%

asarch: That's talking about filesystem type. If it's just something about that partition table, that'd be interesting. You can look at the listed types.

I used freebsd-boot, freebsd-swap and two freebsd-ufs partitions (one for / and the other one for /usr/home)

asarch: gpart show -l ada0 (or whatever) will show partition types.

hm

it says hostname not found

and I cannot ping

I tried turning on nfsd that didn't help or dhclient renew

asarch: FWIW, freebsd-swap should be 516e7cb5-6ecf-11d6-8ff8-00022d09712b, freebsd-zfs would be 516e7cba-6ecf-11d6-8ff8-00022d09712b - not sure where there's a canonical listing.

I don't know why I checked the ethernet cable it was yellow and seemed to work I was able to ping locally

like the router

anyone have an idea? I installed the i386 version bootable and it connected to internet then and got a dhcp lease

I checked the etc/hosts file and adjusted it appropriately nothing special there

ada02 - exbootfs - 1.5G

asarch: Ah, found a list here. There's probably a better one, but: qnx.com/developers/docs/7.0.0/index…/topic/d/diskimage_config_file.html

ada03 - exrootfs - 57G

asarch: Better to show codes.

ada04 - exswapfs - 7.5G

ada05 - exhomefs - 5G7

Those symbols don't convey meaning to me. Try gpart -r show from freebsd boot media

hi mason did you get my question?

kiwichap: Didn't realize it was for me. Reading.

gpart: Unknown command -r

kiwichap: Sounds like DNS is the first thing to check. Do you have a valid /etc/resolv.conf? Can you ping the name server(s) listed thereing?

therein*

asarch: gpart show -r ada0 (for example)

I will go downstairs and check

You mean the 83BD6B9D-7F41-11DC-BE0B-001560B84F0F codes from the partitions?

If yes, that would be for freebsd-boot

516E7CB6-6ECF-11D6-8FF8-00022D09712B for freebsd-ufs for the / and the /usr/home partitions

And 516E7CB5-6ECF-11D6-8FF8-00022D09712B for the freebsd-swap partition

Everything seems to be fine

asarch: Yeah, those.

yes mason thank you

I changed it to google and it worked

8.8.8.8

kiwichap: Ah. The worry there is that now Google harvests everything you do. Better to use your local DNS if you can find it, from some internal router, or from your ISP, or something of the sort.

ok

asarch: Those look right. Might need more context about what's failing.

asarch: The error was about mounting. You're not somehow skipping the actual formatting of the partitions, are you?

The installer says that there is an already FreeBSD system installed in the partition, however, this is actually a blank HDD

asarch: Might be safest to let FreeBSD do the partitioning. There's still something missing in your procedure we're not seeing.

asarch: I'm going to take a wild guess that you're telling it to use the existing partitions, and it's assuming they're already formatted.

thanks good idea it's on the router

Is there a way to force to format the partition?

thank you mason

kiwichap: Sure, have fun!

asarch: That should just happen, although honestly I never use anything but ZFS any more, and at that I do my disk prep by hand.

asarch: I'm not suggesting you do this, but for kicks, an interesting thing that's possible is to spin up a VM and point it to the real disk. You can have your Slackware running FreeBSD as a VM and have it do almost all the work itself from within Slackware.

This is a real PC

I've done this on several occasions. It requires taking care to shift between a config suitable for the VM and a config suitable for the bare hardware.

I just used the Slackware's installation image to create the partitions with its fdisk

Slackware is no longer available

kk, I'd recommend just forgetting about what Slackware set up, and doing a straight install from FreeBSD media then.

Where you are right now in terms of familiarity, you're not helping yourself by deviating from the tried and tested path.

I am planning to use this disk to use FreeBSD, NetBSD, OpenBSD, Debian, Slackware and an Android for x86 installation setup

asarch: rebooting each time?

VM's would be easier

This is an old i5 PC

quit

I would like to use them in real life

I could also load each one of them with Xen from Debian

But that's not the point

Anyway. I think I will have to skip for now the installation of FreeBSD until I find a way to get around this bug

Thank you guys. Thank you very much :-)

anyone know of a good reference on the proper use of the feature test macros (_POSIX_C_SOURCE etc.)? I know all about them myself, but it would be useful to have a link to share

That's the problem with common sense based upon practical experience. We know how things are supposed to work. But not where it is documented.

RhodiumToad, "A POSIX-conforming application should ensure that the feature test macro _POSIX_C_SOURCE is defined before inclusion of any header." pubs.opengroup.org/onlinepubs/009695399/functions/xsh_chap02_02.html

Title: System Interfaces Chapter 2

yes, I know about that

in context, it could be taken to mean "any system header", not just any #include

I would take it as before any <*.h> header inclusion. Not any "*.h" inclusion.

the context here is a library that defines _POSIX_C_SOURCE in its public header file, which is obviously wrong, is fairly easy to explain _why_ it's wrong, but I can't find any good reference or tutorial material explaining it

More standards verbage on it here: pubs.opengroup.org/onlinepubs/000095399/xrat/xsh_chap02.html

Title: Rationale

And here: 9699919799

Oops. Meant to paste: pubs.opengroup.org/onlinepubs/9699919799/functions/V2_chap02.html