lw, exactly :)

lw, I went home again... :s

better not to type wine64 rn

jbo: factorio.exe is sitting right there

lw, ssssh!

i guess this means quality time in the dc with fbsd-14 and zfs

i'm this |-| close to abandoning opnsense and just doing it myself

had to keep killall pkg-static last night to get the damn thing to upgrade because it'd hang, but not a whole lot of visibility through a web UI

jbo: i think i should do something about the wine/wine64/wow thing before i386 actually gets removed as an arch. surely it must be possible to build the 32-bit bits as part of the 64-bit package

lw, I have no idea how upstream works - sorry :/

apparently you have to build it twice, and tell the 32-bit build where the 64-bit build is wiki.winehq.org/Building_Wine#Shared_WoW64 ... this sounds annoying to do in ports

Title: Building Wine - WineHQ Wiki

mainly since ports doesn't do multilib at all, it's either 32-bit or 64-bit

would be nice to fix this in ports itself but... that sounds like a much larger project than i want to take on

fml does freecad now also depend on lang/rust ?!

jbo: your face depends on lang/rust

how do I display all insalled packages that come a particular category, like devel or lang ?

mns i think pkg-query will do it

mns: this is the best way i can find: pkg query '%C %n'|awk '$1 == "shells" {print}'

i thought a pkg-query expression would do this but you can't use %C in an expression

rtprio, lw: thanks. I was trying to do it with pkg-info and was not finding anything in the man page for it

jbo: btw, did you ever hear back from gerald@ about wine?

(just wondering in case that weird mail issue ate the reply)

lw, noperino

i could go for some cabernet sauvignon

there an umozillad firefox in ports like there's an ungoogled chrome?

alepzi: www/librewolf perhaps

also palemoon.org

but i've never used it

i checked palemoon but it's not in ports

lw, the wine+fct situation makes me want to investigate what is necessary to get rid of screen tearing (nvidia GPU). I have noticed it from the day I switched to FreeBSD as my desktop but never bothered to do anything about it. but this is now becoming a mood point

librewolf looks good

jbo: interesting, i've never had screen tearing at all with amdgpu under wayland. i wonder if i had to enable an option for that? (i don't remember doing so)

librewolf works well for my Youngest

alepzi: make sure both the ports and the upstream are up to date with security patches, sometimes these forks can lag behind

jbo: although if you use X, i do vaguely remember there's an xorg.conf option to enable vsync

jbo: check the nvidia driver docs, there's a list of X settings in there somewhere

might be called 'tear free' or something, unless that's the AMD name... i haven't used nvidia for years

lw, I'm doing X + nvidia

lw, yeah I don't even have a xorg.conf

you can create one with the power of vi :-d

hmmmm.... I have /etc/X11/xorg.conf and /usr/local/etc/X11/xorg.conf.d/

lw, actually I do have one at /etc/X11/xorg.conf. that seems to get generated automatically when installing nvidia drivers

lw, I added Option "TearFree" "true" to the "Device" section of xorg.conf and restarted X. the log tells me: "Option TearFree is not used :((

jbo: did you get that from the nvidia manual or from what i just said? if the latter, please bear in mind i have no idea what i'm talking about, i don't use either nvidia or X

Does FreeBSD have the "nvidia-settings" app?

In mine on Linux the screen tearing stuff is called "pipeline" something

lw, my knowledge is inexistent when it comes to graphical/desktop stuff and there seems to be a lot of noise turning up through google

SponiX, would that be ForceCompositionPipline by any chance?

Yeah, that is the one

SponiX, does that go into xorg.conf?

> ForceCompositionPipeline": possible values are "On" or "Off". The NVIDIA X driver can use a composition pipeline to apply X screen transformations and rotations. "ForceCompositionPipeline" can be used to force the use of this pipeline, even when no transformations or rotations are applied to the screen.

On *cough* Linux, I run "sudo nvidia-settings" and enable that under "advanced". Then have the settings program save it to my xorg.conf

SponiX, to answer your question: nvidia-settings does exist on FreeBSD. I'd just have to compile it first

(on linux i had to rearrange my screens with nvidia-settings every reboot)

doesn't it come with the driver?

lw, no x11/nvidia-settings

It will not save and be permanent unless I run it with sudo

ah

jbo: there's a package freshports.org/x11/nvidia-settings

Title: FreshPorts -- x11/nvidia-settings: Display Control Panel for X NVidia driver

time to figure out how to do that without the gui

lw, I know - did I not literally just say that? :D

no, you just said you'd have to compile it...

jbo: I can cat my xorg.conf to a pastebin if needed?

SponiX, that would be lovely - but only if not too much effort/time

I just rarely do graphics stuff so lack of knowledge on my end.

I'm an embedded/headless person

screw eyes!

jbo: oh when i said 'package' i meant precompiled on pkg.freebsd.org. that probably wasn't clear

just to save you the time compiling it

lw, aah, sorry. yeah package != port you're right. apologies.

> Option "metamodes" "nvidia-auto-select +0+0 {ForceCompositionPipeline=On, ForceFullCompositionPipeline=On}"

Given, this is a Fedora Linux xorg.conf with Nvidia. So, a direct copy/paste into FreeBSD is likely to fail ;)

> Option "SLI" "Off" :D

SponiX, so not using the official nvidia drivers? because then they'd be the same

jbo: Oh, I'm using the official Nvidia drivers. Just on Linux

SponiX, should be good then (based on my drivers knowledge)

no more tearing!

thanks SponiX

Glad I could help ;)

back to wasting time with factorio

hah you did run wine64

heh: Approved by: portmgr (blanket, just fix it)

I did

jbo: my question is, why on earth don't they ship with that enabled be default?

Do some folks actually like tearing?

why is there no way to have poudriere testport generate a package you can install. i have to build wine-devel 8 times to test port (4x testport, 4x bulk so i can actually run it)

SponiX, it's not just a "turn of tearing" option. It has more implications

(and unlike every other ports maintainer, i don't have a 32-core CPU to run this on)

lw: I've been known to testport -I it, then rip the pkg out of the jail... not the cleanest thing, but functional

lw: I can loan you like 14 cores and 128G of ram or so - If you need ;)

maybe i should make my testport script do that

I don't even have a FreeBSD vm at this moment though. So it would take me a bit to spin one up

lw, it's all ancient hardware tho.

jbo: you might have a different definition of ancient, my ZFS/VM server is a Ryzen 2700X...

i do having a 12-core Haswell Xeon lying around somewhere but i'm pretty sure even my 8-core desktop is faster than those old things

lol

my 22-Core is Broadwell ;)

2x E5-2690v4 and 2x (something I forgot)

i want to buy something with a decent spec some day (maybe 32-64 cores, 128GB RAM) but i don't know whether to upgrade my desktop or the server

probably makes more sense to upgrade the server since i can run testport and whatever in VMs

lw, I have exactly the same problem right now.

lw, and for me it currently makes more sense to upgrade the desktop :(

lw, I am still thinking about it but I might go Threadripper 7960x

seems very decent price/value

i was looking at Epyc Milan, those are affordable on eBay and i could start with 1 CPU and add a second later

and for a desktop the comparably higher single core is nice

i guess epyc is bad for desktop because the individual cores are a bit slower though

I'm on this X99 2011-3 system until it dies

x99 was nice

Or I die, whichever comes 1st

Mem: 7543M Active, 10G Inact, 1291M Laundry, 9115M Wired, 1104M Buf, 3252M Free

what does "Inact" mean

inactive

...

yes, but what does it *mean* :-P

i thought it meant "allocated, but on the freelist" but according to someone on the mailing list it does not

Title: Memory - FreeBSD Wiki

that's a bit terse... is it like buffer cache pages which are eligible to be freed but could be reused if needed?

(i use ZFS though, so i shouldn't have any buffer cache)

or is it like when application free()s/munmap()s pages they end up as Inact until they're purged and go back to free?

come on tiny freebsd/i386 vm, you can build this port today, i believe in you

apparently giving the i386 vm 8GB of memory was pointless because ARC is limited to 4GB anyway...

weird to think about how we used to worry about how to run ZFS with 2GB RAM

lw sup mate

still having issue to connect in hexchat

same thing ssl handshake timed out

did you ask #hexchat? i'm sure more people there know about hexchat than here

you think is hexchat the problem?

i see in preferences ssl settings disabled

yes, whether to use SSL or not is an application setting

jbo: you have (bug)mail

did you just absue the CC field? :o

> jbo: this is ok to commit

bruh

:D

i believe i "used" it in the correct fashion, to make you aware of the bug!

jbo: i mean it's a final patch, not a wip

unlike the previous patch where the first one hadn't finished testing

also i just noticed bugs.freebsd.org/bugzilla/show_bug.cgi?id=276398 is open... 14.0.0 is already out, but we can't update to that because our setuptools is too old, does it still make sense to update to 13.1?

Title: 276398 – net/py-pychromecast: update to 13.1.0

if 13.1.0 valid upstream realease and everything is fine with the port then sure, why not?

ok, do you want me to cc you on that one too? :-P

you're on thin ice :p

how many ports commits do you think factorio on freebsd is worth?

I have a number in mind but I'm not sure whether it's a good move to communicate that

is it less than zero because no you have no time to do anything except play factorio?

I told my buddy today that you helped me getting factorio to run under FreeBSD and he was like: "oh god no"

lw, in your git commit subject, can you please start with an upper-case letter after the column? Don't update/change the patches you provided. just for the future

emulators/wine-devel: Update to 9.3 (instead of 'update')

after the column?

colon?

oh, colon. ok

english is my third language so idk

i don't normally use upper case letters at all but i'll try to do that

je parlais francais si vous voulez

(but my french typing is bad so... (2nd language))

lw, ports commits are not about personal preferences :p

[00:00:24] Building 26 packages using up to 26 builders

je parle kein Engels

that makes me think that you speak dutch

? it's not a personal preference, it's just how i type, if there's a law that we have to use an uppercase letter there i'll put one there

(a lot of other people seem bad at this though!)

why tf is it building with gcc

does wine not build with llvm?

i was intending to investigate that but apparently it required gcc at least at some point, i didn't touch that

--without-krb5 :o

jbo: out of the last 10,000 ports commits, 2,895 used a colon followed by a lower-case letter, 6,718 used a colon followed by an upper case letter, 387 did not match my regexp. so 30% don't follow this rule

lw, I'm not going to argue about this. No matter what rule or guideline I'll mention everybody (and most definitely you) will find many examples where people did not adhere to that.

sure, i'm just wondering if there's even a point having such guideline when everyone ignores it

jbo you're doing great i believe in you

obviously not your fault unless you were the one who came up with it :-)

(i just find silly rules like this annoying, in general)

hm, i went through all my old portscout emails but there's nothing interesting in there

damn ports maintainers for keeping everything i use up to date

it's almost as if the freebsd ports ecosystem & community is not as bad...

heh, no, net-im/toot is still broken, i just can't do anything about it

lw whats the correct command to update/upgrade freebsd 14.0 in terminal?

warsoul, freebsd.org/releases/14.0R/installation

Title: FreeBSD 14.0-RELEASE Installation Instructions | The FreeBSD Project

jbo pkg update what is for?

warsoul, I do not understand the question

that command what is for?

warsoul, 'pkg update' will fetch the latest package catalogs from the configured repositories (similar to 'apt update' or 'apt-get update')

warsoul: if you're talking about keeping the base system updated, freebsd-update: man.freebsd.org/cgi/man.cgi?freebsd-update

Title: freebsd-update

if you're talking about keeping the packages installed on top of the base system updated, that's what pkg is for, pkg update will fetch the repo info, pkg upgrade will install all available updates

at least for now there's a pretty clear split between the base system and third-party software packages, unlike Linux where the base system is usually just managed directly through the package manager

pkgbase brings FreeBSD more in line with the Linux way of doing things, but it doesn't seem ready for general use yet

i guess that's one way to look at it

more like pkgbase brings it in line with the Solaris way of doing things

what's a 'Linux'?

pkgbase is more like Linux than Solaris

mason: i think it's more like Solaris because it's building packages from a single unified source tree

where Linux builds discrete packages like 'coreutils' and 'diffutils'

i'm poking around sysctl.c a little and I don't see where all the extra categories come from

like on my system I have not just the defaults like kern, hw, vfs, etc. but also ugen1,ugen2,ugen3,security,a few more i can't remember

one of them is... 000 i think. where do i find those definitions for what they do?

Is there any plan to fix the quarterly firefox vaapi error?

fix something? in ports? would you like a pony too

And some candy, yes.

this is the issue where someone backported firefox 123 and forgot to backport the ports it depends on, right?

Yes.

At least, pretty sure that's it.

yeah... good luck.

Mm, no, Firefox 115.8.0esr

That appears to be the only Firefox available to pkg(8) at the moment.

mason, I guess that is a good warning for me who is still running firefox-122.0.1_3,2 okay, but 123 is working through: freshports.org/search.php?query=fir…t=1&casesensitivity=caseinsensitive

rwp: You're not on quarterly ports then?

I believe I am on quarterly.

hm

curious that I only see 115.8.0esr

that freshports page is weird

firefox has 122.0 packages but only for ppc64 and aarch64?

It shows 115.8.0esr as firefox-esr

pkg info firefox | nc termbin.com 9999 here produced termbin.com/yda9

Ah well. Will dig some more in the AM. I should set up a new local Poudriere anyway.

Note that I am running firefox not firefox-esr

But if I search for firefox at this moment there is nothing available. It's a gap now.

freshports.org/www/firefox shows that it has built firefox 123 for latest but not yet for quarterly. It must still be in the build queue somewhere.

Title: FreshPorts -- www/firefox: Web browser based on the browser portion of Mozilla

Firefox 123 has been committed to ports: cgit.freebsd.org/ports/commit/?id=7…b0e524c8d32727681a4324905e350dc433c

Title: ports - FreeBSD ports tree

Firefox 123 is showing up as a failed port: pkg-status.freebsd.org/beefy14/buil…2amd64-quarterly&build=29e7acd4bc33

Title: Poudriere bulk results

I am still struggling to understand and navigate pkg-status.freebsd.org but it's a good exercise to try to work through it.

I think FF 123 has been available for a while I think, pkg info reports that it was built on feb 21

Are you on latest or quarterly? ff 123 has not yet been built for quarterly. freshports.org/www/firefox

Title: FreshPorts -- www/firefox: Web browser based on the browser portion of Mozilla

Firefox 123 for quarterly is still waiting for build dependencies to build. pkg-status.freebsd.org/builds/defau…terly:132amd64:29e7acd4bc33:beefy14

rwp: latest here

any idea what's the "proper" way to restart a NFS server ? I have dozen of clients with many mounts, should I unmount them first?

If you do not unmount, may have to do after the server starts due to "stale file handle" error appearing on clients

... but then my experience is mostly with "soft-intr" mounts

"POSIX.1", "4BSD"... Why do Unix things sometimes use weird naming schemes like that? Why not "POSIX 1" and "BSD 4"?

TuxJobs: BSD prefixes its version numbers because the first one was just "Berkeley Software Distribution" and the second one was the "Second Berkeley Software Distribution" and so on.

They abbreviated it to 2BSD and it stuck that way.

I see. Makes kind of sense.

Before I execute "ifconfig re0 -rxcsum -rxcsum6 -txcsum -txcsum6 -lro -tso -vlanhwtso -vlanhwfilter", I had; options=602019<RXCSUM,VLAN_MTU,VLAN_HWTAGGING,WOL_MAGIC,RXCSUM_IPV6,TXCSUM_IPV6> on my re0 interface

After it, I have: options=2019<RXCSUM,VLAN_MTU,VLAN_HWTAGGING,WOL_MAGIC> and it seemx rxcsum is still there. Any clue?

*it seems

one day i wrote about nginx /robots.txt not matching and in angie it worked right away. 13.2-RELEASE-p9 - nginx version: nginx/1.24.0. deb stable distro works fine nginx version: nginx/1.22.1

j /#jrubberduck

Should one put "USE_JAVA" in a port even when I only need JRE?

I am installing a .jar and need the JAVAJARDIR variable

yet putting USE_JAVA starts installing the whole JDK

Makefile error: "21+" is not a valid value for JAVA_VERSION. It should be one or more of: 8 11 17 18 (with an optional "+" suffix.) Even though in the handbook it's written till 21

hackerman: it might be that USE_JAVA means jdk for build-time, jre for runtime

but, the easiest way to figure this out is to look at the makefiles

meena, wdym by "means"

is there a context in which you can define it, that changes it meaning?

Title: freebsd-ports/Mk/bsd.java.mk at main · freebsd/freebsd-ports · GitHub

pretty sure my intuition there is correct

has anyone come across the issue of moving a main drive to another and having the efi /boot partition stay with the old drive?

i am in a situation where there is a drive that is mounted and for some reason it keeps wanting to use /boot even though the OTHER drive has an etry..

i am trying to get this layout: dpaste.org/oALSV

Title: dpaste/oALSV (Plain Text)

UEFI? If so then you will need to update the EFI VARs. efibootmgr(8)?

to point the /dev/gpt/efiboot0

Needing to manipulate the EFI VARs is one of my biggest complaints against UEFI.

Title: dpaste/fnJpy (Plain Text)

so some how i have to get this to work :)

ahh!

Title: dpaste/LOh5b (Plain Text)

so since boot0002 is (null) at end and boot001 is not

voy4g3r2, Read mason's fine documentation on this topic: wiki.freebsd.org/MasonLoringBliss/ZFSandGELIbyHAND

Title: MasonLoringBliss/ZFSandGELIbyHAND - FreeBSD Wiki

And with that I must afk for a bit.

thank you!

i do not recall doing the efibootmgr steps

voy4g3r2: FWIW, the syntax for efibootmgr is slightly more restrictive in 14.0

lw and I did the UEFI with fallback

we did not do the UEFI with boot variables

lw: thanks for your idea, I was able to update 14.0 into CURRENT and compile drm-61-kmod and x ports! I even got video, though apparently it instantly freezes :D

and i default to, they are the experts.. i was the monkey at keyboard

If you can use the fallback address, that should just work.

yeah for some reason it is going to the "old" drive now.. i want to delete the efi partition and it says it is busy

/boot/efi/efi/freebsd

i do not think that is right.. it should be /boot/efi/freebsd

lets try removing that second efi and see what it does

i should say move the content up a directory level

FWIW, just updated the page rwp linked to note the newer syntax change.

dang it, stills ays that partition is busy

wtf

mason: this is going to sound real dumb, the /dev/gpt/efi0 lines

if i "moved" it do i have to redo it?

so like i have /dev/gpt/efiboot0

crw-r----- 1 root operator 0x8a Feb 29 15:37 efiboot0

which mounts, like your instructins say.. mount -t msdosfs /dev/gpt/efiboot0

is that where this comes into play: dpaste.org/LOh5b ?

Title: dpaste/LOh5b (Plain Text)

i take it <efi> is partition on nda0p1, which then links to the efi/boot/bootx64.efi

voy4g3r2: looking

because for some unknown reason, the ada0 entry in this output, is showing as busy..i can not do a gpart delete -i 1 ada0 dpaste.org/Czg6W

Title: dpaste/Czg6W (Plain Text)

voy4g3r2: So, what you're doing is making a boot variable that lives in your system's UEFI firmware storage. It uses a UUID to find the device with your boot binary, and then the path shows how to get at that from the root of that device.

which, SHOULD be nda0p1

is how i read it

voy4g3r2: Are you trying to delete it from the running system?

the system is running but i am trying to delete ada3 entry NOT the nda0 entry

the nda0 is the new efi

hence the efibootmgr -v "looks" to be telling me it is getting the efi from nda0

voy4g3r2: The boot variables pointing anywhere in particular won't be enough make a partition busy.

Being mounted would make them busy.

Realize that the boot variables are pretty disconnected from the system. From your running system you can see them and set them.

But they don't do anything within FreeBSD.

The only time they're used is when you boot your system, and they're only used to let the UEFI firmware find your bootloader binary, and potentially give it some arguments.

okay, so somethign else must be "making this buys"

So if you're "busy" then either there's something wrong with the device or (more likely) the partition is mounted or part of something like GELI that's using it without it necessarily being mounted.

Yeah.

okay, at least one thing is down

the hunt continues

Something's laying claim to it for some reason.

voy4g3r2: Can you delete it from rescue media / a rescue session?

That's often a good way to guarantee that the running system isn't using it.

i think i tried that already and was having issues

but i am going to try again

this thumbdrive is geting some use these last few days

haha

go figure, the easiest solution si the best

Did it work?

yup

now for the big test

if i reboot, will it boot up using the "right" one

i swore that was done last night..

here we go

voy4g3r2: As a funny little aside, remember that deleting a partition doesn't strictly make the things on the partition go away.

its booting!!!

oh crap

Good good.

Bad bad.

can't open /dev/gpt/efiboot0

so it WAS using that partition

voy4g3r2: Mm. Did you use the same label for the equivalent partition on both disks?

It found a bootloader if you thought it was booting, so the only fix now is to touch up your fstab probably, and/or adjust the labels you're using.

hrm.

i am confused, i thought the efibootmgr -v showed pointing to nda0

but now i have to "recreate" the boot0

that's easy - the basic steps are in my guide

yeah, okay that should be easy

interesting.. your notes mason

you have /dev/gpt/efi0 /boot/efi0 msdosfs rw,late 0 0

but my current fstab

is /dev/gpt/efiboot0 /boot/efi msdos rw 2 2

i guess my question is late and 0 0

I keep two mounts on multi-disk systems.

And I populate them individually.

ah that is the pooling stuff you have at top

I used to do some funny business with mirroring but it was more cute than good.

hrm.. i guess i got to goto rescue mode again

something seems off with the audio/alsa-plugins port. It seems to want to rebuild every time I re pouderire: "alsa-plugins-1.2.7.1_3.pkg: new dependency: multimedia/ffmpeg"

I checked all my make.conf files, I have no options for that port or anything that mentions FFMPEG

well this is just fun

mason: another question.. the steps where you are talking about adding legacy bootcode.. is that required if i do UEFI with boot variables

voy4g3r2: Not at all. You can ignore that.

okay, so somethign is off

i think i need to import my zpool and make the default partition

i just "love" breaking tings

i make the efi partition, move the bootloader.. reboot the machine

and it can not find the dang efi partition

okay, so it boots again.. dumb question

i get the efi error and it is now loaded in "read-only" mode

how would i make it read-write, so i can update my fstab ?

mount -u -w /

(-u ... change allready mounted fs)

Kalten: thank you!

;-)

it works!!!

now my present getting the garbage collected

mason: why do you pick -t msdosfs over msdos in your instructions

i take it both are valid... i do msdos and not msdosfs

voy4g3r2: I tend to start with examples or quick perusals of the docs and use whatever I find works.

or source if there are no docs

that makes sense, this document was VERY helpful

it was more curiousity in my line of questioning and not like.. what is wrong with you, duh it is msdos

if that makes sense