when upgrading, it was noted that running `bectl create` was a good practice, however, I noticed that freebsd-update already invokes bectl. is there any advantage other than a fixed label for running bectl beforehand?

!inspire

Most great people have achieved their greatest success just one step beyond their greatest failure. -Napoleon Hill [zenquotes.io]

zerotime: not really, the freebsd-update behavior is notably relatively new

conventional wisdom can lag behind at times

kevans: thanks, I realize it was introduced in 13.x

i don’t know if this is a freebsd question or not, but: where does the group name “wheel” come from? the common default group name for regular users

I'd be careful about "common default group name for regular users" because wheel is used to allow access to su command;)

The term "big wheel" has been in use to describe important people. In BSD land only people in the wheel group can become root.

The wheel group is NOT the default group name for regular users. It's only for users that can become root.

hmmm interesting

maybe a terminology thing, in mac land, “regular” users are administrator users, and non-admin users are like accounts you’d make for your kids

anyway, i have another more important questoin it turns out

The GID of wheel is group 0. The same number is used in Linux for the root group. Linux does not limit root access to gid 0 like BSD does.

so i run `freebsd-update fetch install`, and it starts doing its thing. But then I get to a point where its basically frozen on the screen and I have to hit ‘q’ to continue. But there is no on screen direction saying this. I NEVER would have figured that out if someone hadn’t told me

can anyone explain what is going on here, why i have to hit q and how do I know when i’m in this situation in the future?

At some point it goes into the less pager and has a ":" in the lower left position of the screen. That's the less prompt. Press q to quit out of it.

why does it do that? rather than just listing every line out like it does for everything else?

and why doesn’t it say something like ‘press q to continue’?

It's just historical. Experienced users know what the less prompt looks like and recognize it.

with lots of output you would miss the first lines. you can try to set PAGER=more ;)

The list of files it is showing is longer than a screen and you can scroll around looking at it. But I don't find it useful to know which files are being updated as it is all of them.

oh well shit, is there a quiet mode for `… fetch install`? that would be a solution

tsoome, Yes! That's a good idea. Set "PAGER=more freebsd-update fetch install" and it will prompt with "--More--" rather than ":".

i agree, i’m doing updates because updates are available, i’m never going to “care” whats being update

Probably you can "freebsd-update fetch install </dev/null" and that would work. Or perhaps "yes q | freebsd-update fetch install". I haven't tried either.

probably worth checking man freebsd-update ;)

oh also, is `freebsd-update` only updating the full release i’m currently on (15)? Or is it going to update me right to freebsd 16 when it comes out?

also regarding the term “big wheel”, this is all I’ve ever known as a “big wheel”

Answers to these and other exciting questions are documented in docs.freebsd.org/en/books/handbook/…e/#updating-upgrading-freebsdupdate

ye, handbook is your friend there.

Basically freebsd-update fetch install only upgrades along minor releases. To upgrade major releases you have to explicitly ask for it. freebsd-update -r 15.1-RELEASE upgrade

You are currently on 15.0 so to upgrade to 15.1 you will have to ask for it explicitly.

Otherwise you just go from 15.0 p1 to p2 to p3 and so on.

oh so it’s apple style, the “tenths” are major releases

In this context... Pretty much.

actually i guess its old apple style, after 20 years, they finally went back to traditional numbering. but anyway

one of these days, when i’m working on learning freebsd, i might actually get around to installing apache and php lol

so i installed windowmaker via desktop-installer, and as usuall… i can’t figure out how to actually launch it, any ideas?

as usually, what i google isn’t working. “wmaker” command not found, and startx kicks me into that windows 3 looking gui

GoSox, startx gives you twm, the standard window manager: <cpcnw.co.uk/twm/twmrc.htm> .

GoSox: if you don't what got installed, you can use `pkg list $pkgname` and then e.g. `grep bin` that

GoSox, first, check your xinitrc file: /usr/local/etc/X11/xinit/xinitrc . It will containt the invocactio of a window manager (probablt twm).

i don’t appear to have a file

which m akes sense, the gui is not auto -launching

well that is me assuming, that desktop-installer is using pkg. you can use `pkg info` to see the installed packages and make sure, that this is the case

ohhhhhhhh i may have installed the wrong thing

oh i’m so dumb

why aren’t i useing a VM to sample all of the DEs instead of installing them on on a real system

oh well too late now

The experience of a DE isn't the same in a VM

how so?

There's a layer of virtualization between, so it feels different. Mouse gestures and hotkeys might not transfer through properly. Graphics and such get filtered through virtual devices with different drivers. It's ok for testing that things work, but if you really want to feel it, bare metal is different.

yeah i guessssss thats fair

turns out `shutdown -h now`, on freebsd, on this mac mini, doesn’t actually power it down. The monitor shuts off and it appears to shut it down, but the power light stays on and the mac runs piping hot until you physically hold the power button

which isn’t really a problem long term, ill be running on an HP rackmount. but its odd on this test machine

GoSox what about using `-p` instead of `-h`? man page explicitly mentions the turning off with `-p`.

on macos, the man page says that -p and -h are the same. and i actually dont’ remember which one i used on freebsd, but i figured since its the same mac hardware, the shutdown results would be the same

AND since i have this mini facing backwards so i can access its ports, i didn’t notice the power light was still on

this state it leaves the computer in, on but no video, is not a state you can put a mac in at all using macos. i’ve never seen a mac do that before. if theres no video, its either booted but asleep, or its off

anyway its not a problem, just an oddity

I'm not sure if the shutdown into halt was ever useful, but it is an option cpus give you, apple probably has -h for compability, but don't want users to really do that, so they changed it to be poweroff instead

hi all.

looking at the man of `cp` command I see in the description of the `-R` option that it behaves differently depending if the source file name ends with a `/` : man.freebsd.org/cgi/man.cgi?cp

so `cp dir1 dir2` should result in `dir2/dir1` ?

i'm trying to install a nextcloud 32 from packages in a fresh jail, but when i run cd /usr/local/www/nextcloud && php occ it's output ends with "Segmentation fault"

is php84 known to be broken?

.oO(insert php joke here)

i'm running 15.0p3 on amd64

i found this "lovely" bug so it looks like i'm not the only one suffering from this insanity

There was a time when the power switch was mechanical, so you had to halt the OS and the press the power button to cut the power

The man page on FreeBSD does NOT say that -h and -p are the same

vkarlsen, I had such a PC, probably 286 or 386. And was literally a switch, not a button.

there even used to be computers that had a power loss interrupt allowing the OS to perform a few 1000 final instructions before the brownout detector triggered

e.g. stop all DMA tranfers

on the DG Nova with core memory some systems just stored the peripheral state and register file

when power was restored the system would just resume

and with core memory you could pretty much continue where you left off, anyways

NAND + DRAM NVDIMMs are basically that again just with better access times (they save the DRAM to NAND flash on power loss)

ant-x: Yeah, I've had a few. I sometimes feel old when talking to people.

my FreeBSD box has its root FS on Optane, which is pretty nice

I wonder how many bugs they'll find in existing software when we have systems that run continuously again (like the core days)

daddoo: doubtful it'll be much new since the limiting factor is almost never having to power things off

could be

I mean "you could be right", not disagreeing with you!

I wonder if/when we'll get to the situation where we lose the hardware difference between what we call "memory" now and what we call "storage".

daddoo: I think the implicit consensus is that that difference is what we call a feature

...except in mainframes/midlevels that use single-level store like AS/400 or whatever IBM is calling it now

and even in those the hardware difference remains

thanks, I didn't realize that about the IBMs, I'm way out of date on their architectures.

daddoo: take a look at rcs.uwaterloo.ca/aurora too! Someone implemented single-level store on FreeBSD

daddoo: it's easy to ignore this on the IBMs if you've never really worked on them, as that's sort of its own parallel universe

thanks for the link, that's cool!

of course, "back in the old days" we had systems without any "RAM" at all ... there were machines that had a few registers, but *all* storage was on a drum.

youtube.com/watch?v=S0HY6Z3aq8A here you go, there's the supposed peak of the "old days" of bona-fide mainframes

at least if you subscribe to their script definition that it's not a true mainframe unless it literally has a main frame which houses the CPU, and channel controllers and so on physically separated

bbl

I've got a copy of Pugh's book on 360/370 systems, the history of their architectures is really interesting

DEC was well-documented also

is there a DEC book you recommend?

"Computer Engineering: A DEC View of Hardware Systems Design", by Bell and Mudge - it's on bitsavers even, bitsavers.org/pdf/dec/_Books/Bell-ComputerEngineering.pdf

it covers up to the VAX 11/780, none of the RISC machines though

is the system 360/370 book anywhere? archive.org only has it as a limited preview

old coworker of mine worked at DEC on the team that made the Alpha

ooh

I never got to use an Alpha, just a MIPS DECStation for a short while though

I think I used sdf.org when they were still on alpha

I haven't run into any DG histories other than "Sould of a New Machine"

there should be a book like Heaths Microprocessor Architectures for every family

see mips run was also quite nice

there are some older textbooks that cover a lot of architectures, too ... I think Gordon Bell wrote one, in fact (not just DEC)

check his wikipedia page (Bell) ... there are two books with "Computer Structures" in the title. THose are interesting too, though not history as such

later all o/

okay so php 8.3 and 8.4 are broken on freebsd 15.0

but 8.2 still works

Hello chaps, I'm trying to get a MacBook trackpad working, would it be OK to post a link to my forum post about it?

Oh well, you can always just ignore it, right? But if anyone's got any advice I would be super grateful!

crest, I am afraid to ask... What's broken about php 8.3 and 8.4? (I'll note those ports are also available in 14.3R too.)

rwp: bugs.freebsd.org/bugzilla/show_bug.cgi?id=291918

So it fails to compile and build, when ZTS (Zend Thread Safe) is enabled. Gotcha!

the default port configuration is broken

the php interpreter crashes on **exit**

iirc ZTS isn't enabled by default for the ports

so it must be something more

i found out when i tried to update my nextcloud to php 8.4

IIRC in PHP upstream ZTS is not enabled by default. Most installations don't require it.

the default is ZTS=off: Force Zend Thread Safety (ZTS) build

on php 8.4

on my test jail php ./occ on fresh uninitalized nextcloud jail just crashes with a segfault at the end of the output

I am curious as to what configuration you are using php running in a threaded environment? (I hope to learn something.)

i just use php

so i ran pkg install -g 'nextcoud*php84'

than added the pgsql and pdo_pgsql modules

but if i enter /usr/local/www/nextcloud

and run php ./occ (the nextcloud cli admin command)

crest: I'm using Nextcloud with PHP 8.4 on FBSD 15.0-R and haven't had any problems at all (also not with ZTS enabled in PHP).

can you can the occ command and does it exit cleanly or does it crash on **exit** with a segfault?

How does Nextcloud run PHP? Does it actually require ZTS mode enabled? I do some searching and it does not seem required.

nextcloud is (mostly) written in php

it doesn't require ZTS

but it does have the problems i described in the configuration shipped as packages

Seems like a lot of pain and suffering can be avoided by not trying to compile with ZTS mode enabled then.

php in ports ships with ZTS disabled at build time

Options :

CGI : on

CLI : on

DEBUG : off

DTRACE : on

EMBED : on

FPM : on

IPV6 : on

LINKTHR : on

MYSQLND : on

NOASLR : off

PCRE : on

PHPDBG : off

ZTS : off

so ZTS is not required to reproduce the problem

That's not bug 291918 referenced earlier then. That was a pre-compile configure time failure bug.

sorry then let me find the right bug

That php zts bug was filed 2025-09-10 so it is aging along and more people are hitting it more recently. :-(