acu: github.com/gokcehan/lf

llua: I would rather say that Windows has been put in a death spiral since Windows 8, and Microsoft just keeps accelerating its death by blindly following the latest tech trends that benefits only the activist investors.

remiliascarlet: not sure how that was relevant to the question being asked, but sure, don't let that stop you from hating.

llua: Obviously responding to this: "it is a dead os"

You didn't asked a question.

which was a response about windows xp

cacm.acm.org/practice/free-and-open…-software-and-other-market-failures has some windows vista foo in it

I won’t remind you how bad Windows Vista was; I will let author Neil Gaiman do that, by quoting what he wrote in his blog on March 31, 2008:

en.wikipedia.org/wiki/Poul-Henning_Kamp fbsd developer, good article

I only heard about Windows being a thing a couple years ago, tried it, got frustrated, and went straight back to Linux.

minix linux open/net/free bsd's ...in 1991 solaris 2.1 for x86, or 1995 solaris 2.5.1 for x86 $600.00 bucks. no compilers

imagine paying $600 for os no compilers

imgur.com/RWy9j4x sun 2/120 1984 $16,900.00

multiuser 3

bill joy he did vi, nfs, bsd! en.wikipedia.org/wiki/Bill_Joy

tcp/ip bsdsockets.

that is before bsdi and open/net/free bsd...

windows a thing ? you cant buy a computer without that os's coming on it, unless it crapple!

the winblows tax!

apple and m$ been assholes from day 1

see the movie ---> Steve Jobs: The Man in the Machine, last half of movie covers some of stupid

The earliest recorded versions of the Double Irish-type BEPS tools are by Apple in the late 1980s

they don't even pay taxes!

foxconn empolyee make $12.00 and apple 300% mark up on i-shit, witch ends up in landfills

apple, nvidia, m$ fighting for most valuable company

AI the new dot com bubble..

remiliascarlet, that IBM for dumping $1 billion into linux in 2000, and then $1billion into linux in 2013, just to fight off m$, of course the redhat purchase is icing on the cake.

thank IBM correction

itsfoss.com/ibm-invest-1-billion-linux example

IBM has been a Linux supporter for a long time. In fact, this is not the first time when IBM has invested in Linux. Back in 2000, IBM had invested $1 billion in Linux with a dedicated team of over 1500 engineers.

rennj: The big irony is that nowadays, Microsoft has among the biggest says within the direction of the Linux kernel.

intel suckered minix into their intel ME ..the BMC. they didnt even give back

yeah m$ purchase of github

but they hoover opensauce into copilot, but not their own code

m$ purchase of linkedin

monopolies,

when m$ bought hotmail.com a fbsd base htmp email server

when m$ bought hotmail.com a fbsd base html email server

ring -3

ring -1 is vm foo, for vmware,virtualbox,xen,qemu

ring -2 en.wikipedia.org/wiki/System_Management_Mode

I am looking forward to Raptor Eng. new boards which (crosses fingers) should be released this or next year, they already have a open source BMC on a fpga, but I think they are going to make a proper one for the next gen.

powerpc foo...

talon raptor

i just want a nice amiga computer/mister in checkmate display...

put a mister in the back, covers couple dozen system up to 486 core in fpga

I bought a few very old PowerPC laptops, since it's the only type of RISC laptops that can be obtained easily.

Installed OpenBSD on the one, and NetBSD on the other.

Linux and FreeBSD are probably going to be too bloated for them though.

morphos

More specifically, it runs on Powerbook G4, iBook G4, Mac mini G4, eMac, Power Mac G5, and Power Mac G4

rennj: Does it use glibc or musl? Because if it's glibc, then it would be a waste of time for me.

its amiga foo

nothing to do with linux or gnu

I said that Linux and FreeBSD are too bloated for macppc, and you said "morphos" like if it were a counter argument like in "actually, you can run this distro".

So I assumed it to be a Linux distro.

nothing to do with linux or freebsd..its amiga like os

So you can play Monkey Island, but not do much else with it. *G,D&R*

i got UNIX utilities on my amiga..

geekgadgets from fred fish is x11,gnu foo on amiga and beos

besides doing amigados and arexx, ibm rexx, i can do gnu foo on my amiga..

like dc,bc,awk,grep,ls and such

besides x11 server if i need it

imgur.com/H2OQ1Qx ffmpeg sounthpark season24 amiga!!!

right from youtube

course its emulation, normal 68k hardware cant handle that

but the software is fine

en.wikipedia.org/wiki/Matthew_Dillon was in the photo with Fred Fish. but hey the amiga sucks right vkarlsen ?

lets boot up the amiga, i can do nc termbin.com 9999 right from the bin/ dir...so vkarlsen can figure it out...

termbin.com/zqa8 unix enough for you

imgur.com/8sWNmx5 and screensht

rennj: I see ed in the list. I approve.

i got diff dd du echo env

i give a shit if you approve

sound like a moron about money island

sound like a moron about monkey island

you have no clue

There's no need to get so worked up over a joke :D

oh its a joke now

didnt sound like it

Did to me

<vkarlsen> So you can play Monkey Island, but not do much else with it. *G,D&R*

some joke

Note the last part of my message

?

gobbley gook

The grin, run and duck part?

the amiga outlived tons of boxes and is still rocking to day. apollo-computer.com

find a sun box that runs ffmpeg

good luck

sun ultra 45 /sparc or hp c8000 pa-risc... i doubt i could handle the stuff my amiga can do..let alone compile modern software.

sgi tezro / mips

dead

sparc/pa-risc/mips but the amiga keeps on jugging along.

Hello, all. This script <cgit.freebsd.org/ports/tree/emulators/wine/files/pkg32.sh> invokes `pkg' with `-o ABI_FILE <path>`, whereas I do not see an ABI_FILE option in pkg.conf(5). But it has ABI, which may well be the same thing as ABI_FILE. Is it an undocumented synonym?

Oh, -o ABI specifies the ABI directly as a string. So, ABI_FILE is an undocumented option of pkg?

I've got a bridge on my host for VNET jails. Can anyone think of a reason why at some point, packets (replies) to a jail address would be visible on the (host) igb0 in tcpdump, and would *not* be visible on the igb0bridge device in tcpdump?

I haven't messed with bridges in ages, but I was under the impression that if a bridge contains the physical interface, then packets would have to be visible on both or neither, and this is not the case now

More specifically, if I ping6 out from the jail, the echo request is seen on igb0bridge. But the echo reply appears on igb0, not igb0bridge.

ant-x: github.com/freebsd/pkg/blob/main/libpkg/pkg_elf.c#L837

I am sorry -- got disconnected because of a pooor internet connection, and then forwarded to #freebsd-irc.

ant-x: github.com/freebsd/pkg/blob/main/libpkg/pkg_elf.c#L837

vkarlsen, Yes, I have seen that file, and still see your URL in my clien't IRC log. I just wanted to cofirm that ABI_FILE is an /undocumented/ option of pkg configuration.

In other words, ABI_FILE is not mentioned in <man.freebsd.org/cgi/man.cgi?pkg.conf(5)>.

Who should I bug to add makecontext(3) variant that takes va_list argument instead of ellipsis?

cracauer, ping

FreeBSD has a concept of ABI: an string identifier of OS version and platrform, e.g.: FreeBSD:14:i386 . As fix to this bug: <bugs.freebsd.org/bugzilla/show_bug.cgi?id=278525>, the script for intalling 32-bit Wine was modified to take the ABI from the file /usr/lib32/libc.so.7 instead of from the output of `pkg config ABI | sed s/amd64/i386/`. When do these methods return different results on AMD64 hardware?

The latest pourdriere-devel is a game changer with PKG_NO_VERSION_FOR_DEPS=yes. No more unecessary builds! w00t! freebsd/freebsd-ports eef5ff6

yeah that was an interesting changelog to read the other day. I have yet to see how that feature pans out for production runs. probably not going to touch it for a while outside of testport

ant-x: I wouldn't have thought they would

oh

looking at the context, I think the point is that overriding ABI is wrong and one should be using ABI_FILE instead

less fragile

kevans, but is not ABI_FILE just another (undocumented in pkg.conf) way to override the native ABI?

kevans, You say it is less fragile, but I say it is /more/ fragile: for now it depends on the existence of a certain file at a certain location.

Is noitthe following a safe version:

ABI=FreeBSD:$(./freebsd-version | sed 's/\(^[0-9]*\).*/\1/g'):i386

^ omit the ./ from the invocation of freebsd-version, it was a stub script.

ant-x: ABI derivation always relies on a file at a certain location

What I mean it: why not read the OS major version and then generate the ABI string ourselves?

kevans, but there is no such file in my FreeBSD 14.1, so the derivation does not work.

then you did not install lib32 and the 32-bit wine won't work anyways

kevans, but pkg should have taken care of it, no?

unless we bumped libc.so soversion, but i'm pretty sure we didn't

lib32 is part of the base install, nothing pkg

How come it is not there in my FreeBSD 14.1? Did I neglect to select it during setup?

yes

OK, thank you.

Back to my original proposal: since 32-bit Wine is for running i386 Windows programs, what is wrong with generating the ABI string my way, as shown above? -- get the currect major version of the OS, and append `i386'?

kevans, Is there a way to install lib32 without reinstalling the entire OS?

I hope the proper solution for running 32bit windows programs with 64bit wine is ready soon (that pkg32 hack additionally installs a 32bit wine and has some wrapper script to select and run the system installed 64bit wine or the 32bit one installed in HOME)

nimaje, Yes, it is. And it disagrees with the official Wine Qucik Start Guide, since wine tried to run the 32-bit version, and for x64 one has to execute wine-64.

In the mean time, can you please help me use what I got from the quarterly release? First, I think I have to get lib32 isntalled.

I think all that is required is extracting lib32.txz from download.freebsd.org/ftp/releases/amd64/14.1-RELEASE to / but not sure on that

nimaje, Does FreeBSD maintain some index of installed libraries and their versions and locations to help with linking?

maybe using freebsd-update would work, if you get it to accept that the lib32 component should be installed too

But then I have /really/ to update my OS, e.g. to 15.x ?

jbo: pong

cracauer, you're a rust person now, right?

No.

no the idea was, that it thinks some component should be installed and is missing and then fetches and installs that

cracauer, aye

cracauer, don't worry - I have a 2nd issue to bother you about :p

I want to check out the macro systems, but didn't have time yet.

cracauer, I remember you saying "I'm a rust developer now I guess" or something along those lines.

That was in the heat of the moment of not having reusable hashtables in C :-)

cracauer, hah

cracauer, so I haven an I226-LM NIC on my new mainboard. It generally works. the igc driver attached to it and I have been using it for months now. However, the interface only gives me 100 mbps speed (consistently) although ifconfig reports 1000base-T.

tried different cables, different switches, iperf to different hosts etc. it's always pretty much exactly 100 Mbps

who do I talk to about this? or where do I assign a PR for?

Isn't that a 2.5 Gb NIC?

yeah

but it's on a 1G network

Is the problem there when you connect it point to point without your switch?

yep

I'd mail the -net mailing list.

-net? is that net@ ?

errr, hold sec

freebsd-net@

ack - thank you!

I'll re-run tests and then ask there.

I have just been doing 100mbps the past four months :p

nimaje, I understood your idea, and wanted to know which command to use wiht freebsd-update: not `upgrade', but `fetch'?

cracauer, You still need generic hashtables in C&

& -> ?

No, I used binary search. It was more of a general complaint about reusabilty of algorithms in C.

But if you know a good one I'm all ears.

yes, fetch and install, but no idea if it works

Imagine having templates

C++ **cough**

cracauer, apparently these are all the crates necessary for a console-only CSV viewer: bz-attachments.freebsd.org/attachment.cgi?id=251523

cracauer, That is a good point: I like making generic reusable algorithms in C, even containers. They are not in stdlib, but can be written once and reused many times. There are several approcaches to this.

nimaje, Will try.

cracauer, I have an (unfinished) concept if generic dynamic arrays: <sr.ht/~shepton/array.c>

cracauer, I cannot remember the generic container library that I especially liked, but there are many, e.g.: <github.com/JacksonAllan/CC>, <github.com/stclib/STC>

Thanks.

cracauer, I will be back if recall the one I /did/ like. And of course it great fun writing them yourself.

jbo, cracauer: "Object-oriented programming in C" : <archive.org/details/ObjectOrientedANSI-C_1.3c>

very familiar with that.

plenty of OO C here: git.ugfx.io/uGFX/ugfx

uGFX in C. Will it work on desktops?

ant-x, yes. and unlike other embedded GUI systems it runs natively on a desktop and not via an emulator or simulator

ant-x, queue implementations: git.ugfx.io/uGFX/ugfx/src/branch/master/src/gqueue/gqueue.h

That rust dependency list looks impressive. Especially if you consider how many are version < 1.0.

yeah these days releasing often seems to be the goal, not releasing quality.

wazzup cocks installing openbsd 6.3 back on my laptop cos its better than the latest 7.5 xD

psionic: and that has something to do with FreeBSD ?

SponiX: Nope. Only something to do with c**ks, apparently.

That's probably enough of that.

Yes going down in quality xD

I went with obsd because fbsd nbsd wasnt working fine for the purpose at all so you are < 1

But rather saying its fagfox probably more trashed

I aint wanna bother with compiling old firefox on new systems so better going back

wooh, who knew my machine would be 10x quieter if it didnt have 90 million crc errors every read

wabkia: seems legit

back to the update woes :X lol

I think we should go back to http1 and gopher and stay there it was good that way it was

big agree

freebsd 11 was good i am not sure what is gained by updating

I need to narrow down why my desktop in 14 takes a pause for about 30 seconds every so often. It did not do that in 13.

I suppose I could boot back into 13 for a while and see if it is gone but it's unpredictable.

rwp you checked dmesg?

storage errors?

cron? hourly scrub?

Nothing of note in dmesg or /var/log/messages no zfs errors no zpool scrub running not coalesced with any cron run that I can determine. But thanks for the brainstorming!

I wasn't really expecting to debug it at this moment. I was just commiserating with the theme that these upgrades are not improving things.

It does seem to be more tied to doing something in Firefox and Firefox calls fsync() a lot making me suspicious that something that was previously asynchronous with zfs is now synchronous and causing the pause. But that's a guess in the dark by a lot.

ah yeah i hear you. two years ago this upgrade from 12.2 - 12.3 failed. i am finally deciding to work on it. i didnt have my server for a while, its tied to some creative work i was not doing. but now i want to do it and it should be a simple upgrade. too bad it doesnt want to boot after a minor revision lol

rwp its possible zfs needs to be told to update, thats an odd one though!

I am holding off upgrading the zpool because once I do that I can't boot back to the previous 13 anymore.

I shouldn't need to upgrade the zpool and it should just mean I can't take advantage of the new features. Features that I have never used before.

I have 16 GB of ram and right now it looks like I am sitting with 4GB in userland use and about 4GB in file system buffer cache, just eye-balling top statistics. So plenty of RAM not yet doing any work for me. I rebooted yesterday.

30 second pause definitely feels like something is being accessed, and timing out.

I fear that it will be either 1) kernel change that now causes this behavior 2) zfs change now causing this behavior

I haven't put a stopwatch to the pause but it's at least 15 seconds and less than 45 seconds somewhere. That does feel like a timeout of some sort. And if I wait then it pops back into responsiveness again.

thats rough =\ i def feel ya there.

I have 214 active snapshots but I run zfs-autosnap and have had 356139 snapshots in the zpool history which makes me wonder if something in there is now bogging down.

ZIL?

is there a man page that documents control codes? e.g., ctrl-c

Big upgrades like 13 to 14 could introduce problems in lots of places. It might be the graphics driver which is taking a pause and of course I am using X so the pause could be solely interactive through the graphics system and if I were on an ssh login I might not be experiencing it.

I suspected the ZIL with Firefox + all of those fsync() calls but I don't know. ZFS itself seems happy by looking at the top header stats. And I am usually not writing huge data on my desktop when I notice the problem.

markmcb, Control-C is handled by the tty driver. Look at "man stty" as a first start.

Run "stty -a" and look and see that intr=^C which is handled by the tty driver which then sends SIGINT to the foreground process group.

rwp, thanks!

Honestly the best source of documentation for all of this is Advanced Programming in the UNIX Environment by the late W. Richard Stevens now maintained by Stephen A. Rago. It is a large tome in the latest edition. It is an excellent resource for these things.

I can't recommend it highly enough. There are other resources too but if I have only one recommendation then it is that one.

markmcb, Secondly look at "man ascii" which documents the list of characters which includes the list of control codes in there too.

do i need opensolaris_load="YES" in addition to zfs_load="YES" in loader.conf?

maybe this changed from 12.1/2 to 12.3? its a new kernel, after all.

but i thought opensolaris was for dev versions of zfs

I know nothing about opensolaris_load="YES" and don't have it here and so "need" would be a definite no. But no idea if it does something for you or not.

zpool version returns "zfs-2.2.4-FreeBSD_g256659204" "zfs-kmod-2.2.4-FreeBSD_g256659204" here.

my version doesnt return lol, or im getting the property wrong

"zpool version"

i guess that didn't exist in 2017 :)

I am on 14 now. And that has switched to the OpenZFS upstream source.

looks like that happened in 13, but im just going from 12.2 to 12.3 hmmmmm

maybe i just add in the opensolaris_load to loader.conf if its not going to wipe the pool

zdb reports version: 5000

As I recall the feature version was converted to 5000 as somewhat of an infinitely large version when they switched to feature names.

hmmm

Is this something that would relate to your environment? forums.freebsd.org/threads/zfs-upgrade-from-version-28-to-5000.59332

well i am worried if i zpool update i wont be able to boot into the old environment or the new one lol. gonna figure it out without an update for now. i have a feeling the updated zfs is in limbo

Makes sense to me.

I have lost sync on your issues. Could you summarize them again?

me??

its all good, appreciate the help :) i am just trying to update from 12.2 to 12.3 and when it tries to boot after freebsd-update install it gives me a mountroot> prompt

and dont worry, i think ill figure it out at some point

I don't recall hitting that problem when I upgraded from 12.2 to 12.3. I am sure I would have remembered.

Have you updated the boot code? Something like: gpart bootcode -b /boot/pmbr -p /boot/gptzfsboot -i 1 ada0

i did try that just now, yes :)

gonna try updating loader.conf with opensolaris, maybe if i grepped the update i would see something about that

kevans: I am installing the 32-bit libc via bsdinstall, according to this answer: <unix.stackexchange.com/a/563156>. I hope it does not format my HDD

unfortunately adding opensolaris_load and _enable didnt help =\ sigh

solaris lol grandpa ok

ok so i think the problem might be that im gpt/bios booting and i need to efi boot

JFTR but I am still Legacy BIOS booting.

=\ yeah i dont even know if i have EFI. its a pretty old system, but still x64

seems like i could go this route

fetching 68k files lmao. sheeesh

this cant be real :( "///usr/include/c++/v1/__string exists but is not a directory"

hmmm ok, i am able to boot into 13.3 no problem, but only if i select the boot env from the boot menu? how do i get it to zroot/ROOT/default?

seems like i should do this? bugs.freebsd.org/bugzilla/show_bug.cgi?id=273661#c31

separate issue ig

wabkia, Which boot environment works?

The normal plan, which sometimes gets disrupted, is that zroot/ROOT/default is like in a version control system the main trunk of the file system.

Every time freebsd-update changes things it will make a new snapshot and clone a new fs to be a new Boot Environment causing those to be like forks off the main trunk.