a red box with listings of hex numbers means that it has found errors.

RhodiumToad: LxGHTNxNG: 0x0.st/HPqm.jpg

Yes. I guess it has found some errors.

It seems it has found many errors.

memory errors

yeah. that's not good.

have you adjusted any of the bios timing settings, for overclocking or whatever?

I am doing a zfs send and receive of a snapshot between two hosts connected by 10G ether. The process is printing out a line once a second: 17:03:27 188K rz2_pool/homes@2023_04_21 is this zfs telling me it's sending 1Mbit per second? Am I interpreting that right?

No. Nothing.

ddr3 @ 1846MHz 6-6-6-15!!!

RhodiumToad: E-series pentium dual core. I really doubt it.

oh, no - 923=461*2

what voltage do u use for ram?

may be set 1.6 or even 1.7

VVD: I do not know really. I'm a newbie.

pr-asadi: Try removing one of the memory sticks and rerunning memtest86plus.

check bios settings for ram/orverclock/etc

pr-asadi: From whom did you get this system?

it can be 1 module 4GB

VVD: E-series pentium dual core. I really doubt overclocking was on the vendor's mind.

LxGHTNxNG: I do not know. My father bought it when I was 5. And now I'm 20.

eh, I had a desktop of similar age and it had a load of bios settings for tweaking

VVD: I will check those settings, but I'm sure I have not changed them.

pr-asadi: Can you identify the RAM configuration in the machine? It may have just one stick, or it may have two.

it can't be quite that old, from the intel specs; they show the CPU as being introduced in 2010

still very old though

If it has two, remove one and retry memtest.

LxGHTNxNG: It is a two channel. I have two memory sticks. Each 2GB.

RhodiumToad: That's 13. Also remember LGA775 played host to Pentium 4/D as well as Core 2.

pr-asadi: Remove one of them and retry the memory test.

LxGHTNxNG: OK. Let me shut down the PC.

I have older one - E6750 Conroe (65nm), E5700 - Wolfdale (45nm)

I think it's going to be the one closer to the CPU.

2GB will leave you a bit tight on ram, you'd probably want to replace it

RhodiumToad: Yes, but for identifying the faulty stick this is fine.

yup

maybe it can work fine with higher voltage, or higher timings

VVD: shut up for one minute

lol

RhodiumToad: It was 2GB before, I bought another 2GB RAM to have 4GB RAM.

pr-asadi: Which stick have you removed? The one you installed at first may be faulty.

LxGHTNxNG: The new one. I mean the one which I bought it myself.

RhodiumToad: Timeframe clarification: Pentium D 945 was RTM Q3'06. This machine has likely seen a CPU upgrade.

if pr's assertion of the machine being bought 15 years ago is correct.

pr-asadi: Right. Rerun memtest and let us know if it shows errors.

LxGHTNxNG: It shows errors.

:(

this soon? Power the box down, yank the DIMM, and put in the one you just removed.

OK.

Attach a sticky note to the faulty DIMM saying "faulty - 30,000+ errors in Memtest86+" in English and your native language.

What I want you to do is install the new DIMM in the slot the old DIMM was in. If memtest shows errors, the fault is with the mobo, not the DIMM.

LxGHTNxNG: I'm sorry, what is DIMM?

Dual Inline Memory Module.

LxGHTNxNG: Exchanging the slots?

colloquially "memory stick"

There is no errors yet.

Okay, wait until it wraps around

OK. I'm gonna wait. After it is done, I will change the places.

The old memory goes to new memory's place and the new one goes to old memory's place.

Yes

If you get errors at the same location, it may be the motherboard. If you get errors at a lower location, it's likely the old memory.

(or you can do this without the old memory, and just put the new module where the old one was; it should still show no errors. if it shows errors then the MB might be broken.)

LxGHTNxNG: I have put the new memory in old memory's place.

Still no errors.

okay.

let it run to the full 2GB.

Sure.

any errors?

if not, the DIMM you just removed is the faulty one. you're going to want to stick it with the sticky note i mentioned earlier.

Not yet.

No errors still.

The brands are different. No sticky note is needed. The old one is Apacer, the new one is Kingston.

If there's still none after it's done the whole 2GB in all its different patterns, then yeah, the DIMM is kaput.

Still need to sticky it so you do not accidentally sell it.

and anyone who happens upon it knows it's junk.

I do not have sticky note around, but I can mark it. And I won't sell it.

DIMM is the old one right?

DIMM is a generic term for any dual inline memory module. The one you want to mark as broken is the one that is broken.

Well. Should I not put the old one into the place of new memory?

You told me to do it few minutes ago.

If you want? But it'll just have errors. I would not try to use it.

The one not in the machine, mark it as broken. No further testing is required

Hmm. So I should buy another one.

Yes. But for now, you can get by on just the one.

LxGHTNxNG: What would happen if I use it? (I'm not really going to use it, just asking to know)

LxGHTNxNG: Of course I can. I used FreeBSD with 2GB RAM for 2 years.

pr-asadi: It would experience the same errors that it did 30 minutes ago that caused the hash to be different every time.

LxGHTNxNG: Well. It won't be harmful right? Just hash stuff?

I did not know about these stuff. Thank you all, LxGHTNxNG VVD RhodiumToad.

It will be harmful. Programs will behave in strange ways.

Things Just Won't Work.

Or they'll return garbage.

LxGHTNxNG: Hmm. Thank you LxGHTNxNG. I did not know about these things. Thank you.

You all helped me a lot.

LxGHTNxNG: VVD: RhodiumToad: Thank you all.

if only I had a lot of money so I could snag a skylake (or later) system for this kind soul

LxGHTNxNG: One more question. I'm re-doing the checksum on the bad files. They are still saying "FAILED". This is because some things about them are produced/made by the old RAM?

Probably, yes.

Delete the files, and retry.

OK. I'm gonna download again.

LxGHTNxNG: It is alright. Thank you for your time.

i have never installed freebsd before. Can I expect wifi out of the box on the memstick.img? will wpa_supllicant be there? thankyou

morpho: you can use wpa_supplicant wifi on my dell laptop driver works for wifi

its going to depend on the hardware you have

my wireless chipset is in wiki.freebsd.org/WiFi/Iwlwifi/Chipsets

Title: WiFi/Iwlwifi/Chipsets - FreeBSD Wiki

its a thinkpad x270

Show me a person and I will show you their crime. I don't care if they're the most moral person we both know

hi! im trying to startx freebsd 13 on an intel, however i get the error: vm_fault: xauth: .serverauth does not exist.. pager read error, pid XXX (Xorg).. xinit giving up.. xinit: unable to connect to X server: Connection refused.. xinit: server error

any idea how to solve this ?

its a new core i5

i tried to add i915kms to rc.conf but didnt help

trying to startx with root gives same error

its a system76 laptop.. that was expected to work flawlessly

hernan: where exactly did you see those errors?

pager read errors aren't expected - any indication of hardware trouble?

RhodiumToad: dont know, i installed a 13.1 and trying to startx i got those errors

trying to install 13.2 now

but those errors appear after "startx" command

and i had exec i3 in .xinitrc

with i3 installed

thanks for the help RhodiumToad

i will need to sleep now

will try again tomorrow

thanks so much =)

<04:28 <hernan> its a system76 laptop.. that was expected to work flawlessly ⬅️ maybe by now the WiFi works

Hmm, I wonder if this could be a reasonable alternative to poudriered (which seems to have stalled?)...

Title: GitHub - Nukesor/pueue: Manage your shell commands.

pr-asadi, there is a small probability that your faulty memory stick can work if you manyally increase voltage and/or timings

hello. Is there no Chrom{e,ium} pkg for arm64 in 13.2-RELEASE? Firefox broke on the update from 13.1 (segfaults every 5min or so) and I'm looking for another browser. I was surprised not to find them in the pkgs and I'm now left without a serious browser in my Raspberry Pi

alternatively, does anyone suggest another browser that isn't Firefox or Chromium that can do banking with relatively little hassle?

i like qute

oof browsing the web on a pi sounds horrible :o

it shouldnt be, but the js devs ruined it for everyone

and the possible py-script devs will make it 10 times worse

greetings, how are jails worked with now-a-days? I see iocage ( and perhaps a few others ) have been abandoned. Is it recommended to use bsdinstall or the older methods using system src and compilation?

iocage abandoned? i use bastille, but i thought iocage and ezjail were around

rustyaxe: browsing with full Firefox on the Pi4 was actually quite usable in 13.1-p7, comparable to a 2012~ mid-hi range laptop. I think it was some powerd magic, because no Linux distro ever reached that sort of performance. But 13.2 upgrade broke the browser so I'm sort of annoyed. Might just roll back and sit around until 13.2 matures a bit :P

and that including the whole Javascript framework madness nowadasys, and even (gasp!) Office365

idk my high end 2012 vintage laptop (dell latitude) is becoming a relic :o

ayy I still have my latitude from 2006 in my parents' place. It can do my banking and even online shopping so I take it's still good!

it's probably not powerd, no.

Does anyone know 13/stable fails compiling with " cc: error: no such file or directory: '/usr/src/sys/amd64/linux32/linux32_support.S' " -- there's the file but with a lowercase "s"

Onepamopa: Might be worth a bug report.

by the time they read it, someone would've committed a fix already

Onepamopa: I repeated this on EFnet in a developer-heavy channel.

Onepamopa: It's being looked at now.

Onepamopa: They said it compiles as-is, with .s, as of commit 517ccb7c8061ace6565fb27f8f0de7bb9305e853, so maybe update?

well, maybe they should checkout a fresh source tree and try again ?

or do they need a screen shot ....

Ah, they tried from main, and they're trying again with stable. I assumed wrong.

1. git clone the src tree; 2. git checkout stable/13 and then try to compile ....

the "master" tree does compile, but stable/13 doesn't

feel free to reproduce

Onepamopa: I just run releases. Are you compiling from an earlier stable/13?

from 13-stable to 13-stable

Looking for some quick suggestions: I have a directory of packages (*.pkg files). I want to delete all but the latest version of each package (e.g. Given files curl-7.1.0.pkg, curl-7.2.0.pkg, curl-7.3.0.pkg...I want to delete all EXCEPT curl-7.3.0.pkg). What different approaches can be taken to reliably accomplish this?

VVD: I have to increase the voltage in BIOS?

_xor: Programmatically identify each base name, then sort of a list of each and keep the head or tail or however you want to do it.

Originally, I was going to sort the filenames, remove non-duplicates, and then just delete all but the most recent one from that list. BUT, I'm not sure which character to split the packages on, since '-' isn't reliable.

Onepamopa: What's the top commit?

mason do tell them to do git checkout stable/13 and compile

Right, that's what I was going to do, but there are cases where you could delete packages that were meant to be kept.

Onepamopa: Might be easiest for you to hop on EFnet/#bsdmips to connect directly.

Gah, I just had an easy way to do it in my head but I lost it.

mason does efnet have a webclient ?

Onepamopa: Interesting, so I learned about ci.freebsd.org/tinderbox thanks to your making a noise.

Title: FreeBSD Tinderbox

_xor: to make sure you parse name/version correctly, try using `pkg query --file this.pkg <pattern>`

Nice

Onepamopa: Hm, yeah, looks like these IS one. I'd have thought not: irchelp.org/networks/efnet/servers.html

Title: IRCHelp.org — EFNet Server List

uhm.. "We couldn't connect to that server :(Unknown error"

Onepamopa: git rev-parse origin/stable/13

(as stable/13 is a branch and not really a commit which was asked about)

517......853

ok, so it builds for everyone except you, it seems :D

(cannot copy, that machine is on another monitor just terminal)

yuripv how come? I just clone the src.git repo in /usr/src then go into /usr/src and do git checkout stable/13

***nothing else***

then just make -j8 buildkernel

and it fails after a minute or so

yuripv: Yeah, looking at that very man page right now :D

yuripv: Though I was kinda hoping there was a rope-equivalent to pkg-check :/

*repo-equivalent

Then I'd just build the repo with multiple versions of the same package, run a check on it to generate a list of duplicate dependencies listed, and then delete the old ones.

...and re-run pkg-repo (yeah yeah, inefficient).

yuripv if you're telling me that somehow I'm getting a completely different source from git then... I've no words.

Onepamopa: The suggestion from #bsdmips was that maybe you have something local or environmental that's breaking things.

WITHOUT_CLEAN=1 or KERNFAST=1 or something

yes, I have those, but I'm running a completely "clean" "src" folder (removed src and checked it out from git)

It build for the person and for Jenkins, so it might be worth identifying what's different.

so it builds everything from 0 without having anything changed or "local" or what not

Onepamopa: i'm not telling you anything other than it builds on tinderbox; you also need to clean the obj directory

Worth a bug then, noting the commit along stable/13 that you're running as you build.

yuripv where's the obj directory?

/usr/obj, by default

oh

Possible for it to be elsewhere though.

it's in /usr/obj

can I safely rm -Rf everything in it ?

Yes, it'll be rebuilt.

It's empty by default.

ok I'll remove the src folder and everything in the obj folder and will try to rebuild again..

Onepamopa: You shouldn't need to remove source.

just in case ...

FWIW, >>> Kernel(s) GENERIC built in 196 seconds, ncpu: 20, make -j20

That's...quite nice.

$ git rev-parse --short HEAD 517ccb7c8061

_xor: that was for Onepamopa :D

yeah yeah I've got a 48 core cpu too, just not on *that* machine ;)

for now everything looks promising, no error...

so the "cached" objs were the issue

Greetings. I have FreeBSD-13.2-RELEASE, there is no telegram-desktpo package. I'm using pkg(8).

telegram-desktop*

pr-asadi: freshports.org/net-im/telegram-desktop

Title: FreshPorts -- net-im/telegram-desktop: Telegram Desktop messaging app

bugs.freebsd.org/bugzilla/buglist.c…ktop&short_desc_type=allwordssubstr nobody has complained yet

Title: Bug List

pr-asadi: looks like it's in quarterly, but not in latest

Hmm. Thank you.

I should switch to quarterly then.

pr-asadi: might be worth building yourself with MAKE_JOBS_UNSAFE=yes and see if that fixes it

meena: Sure. What does MAKE_JOBS_UNSAFE do?

builds using single make job

that's going to take (a lot) longer than usual, but can fix some failing ports

pr-asadi: what yuripv said.

to elaborate on the name: it declares that `make -j` is unsafe (with any value other than 1)

Is 0 "normalized" to 1?

if (*p != '\0' || opts.maxJobs < 1) { (void)fprintf(stderr, "%s: illegal argument to -j -- must be positive integer!\n"

now we know!

phttbbt

pr-asadi, yes. Can you show text from both memory sticks?

and we can promptly forget it

:D yeah sorry, i don't know why looking in the code seemed easier than just trying to run `make -j0`

I was very briefly wondering that, but then I figured, eh

Hey, I have a question. I want to install nextcloud with postgres. And, I have postgresql15-{client,server} on my machine, but nextcloud built from ports uses php81-pdo_pgsql, which in turn depends on postgresql13. Can I somehow built it against 15?

Emru: You can set DEFAULT_VERSIONS= pgsql=15 in make.conf

vkarlsen: thanks!

CONFIGURE_ARGS+= --with-pdo-pgsql=${LOCALBASE}

that is not very helpful

i think

but what do I know

Hi, after may hours struggling with a connection to a pptp server I found the problem was my firewall (pf). Now I need help to allow connecting to the VPN without disabling pf.

The interface created by the VPN is ng0, can I add it to pf.conf and allow all traffic pass through it?

The rule would be: pass out on ng0 ... # Adjust for log, state, protocol, etc

... should also add a rule to allow ingress traffic _carefully_ if the connection would be initiated from outside to estabish handshake

s/estabish/establish/

sorry I was fiddling with pf and part of the conversation was lost

gone again 🤷🏽‍♂️🤷🏽‍♂️

ttzt

oops, wrong focus

meh, parv's emoji looks decomposed, same as meena's; wonder if it's irssi built without utf8proc or something else