hi people!

hello

it's been a long time

I am a member of Turkish local hackerspace in Telegram also

And also I'm admin there

lovely

gry if not a problem where are you from ?

Australia

Nice Country

i've not been to Turkey, and i've not seen members of local unix user groups, bit shy of new places so

yes, it is great so far

Do you have a server located in your country ?

I have a website which is "shared hosting" but not a server located in a dataceenter

I was running FreeBSD in GCP

hwpplayer1: i got a vm in south australia, from networkpresence

datacenter sort of thing gives me more control, but is more expensive

true

i do have a static ip at home so it could be used for some services shared with a group of friends, but nothing production level

understood

let me search networkpresence

 

 

sorry for copy paste

"... awesome!! They solved the problem before I had finished explaining it." :)

this is a bad commit message for this change bugs.freebsd.org/bugzilla/show_bug.cgi?id=271243

Title: 271243 – sysutils/tuptime: Fix cron file

A bit confusing, yes. I had to look at the patch to understand what it meant.

meena: A more "correct" way to give feedback like that is to find the mail in the appopriate commits mailing list and grab the mbox file, then open it in your MUA and responding to the commiter that way.

ok, so /home, or /usr/home? ;]

/home

or third option, /usr/home with a symlink from /home

but you have to create it yourself prior to creating any users

the /home -> /usr/home symlink is evil

haha!

i do that often on clusters ;]

As an example, I committed something with a bit too little context, which got filed as lists.freebsd.org/archives/dev-commits-doc-all/2023-May/002008.html - and then did the above (except that I replied to the mailing list).

Title: git: e7be269af4 - main - Added missing -u flag to pdbedit

cgit.freebsd.org/src/tree/usr.sbin/pw/pw_user.c#n115 /usr/home according to this

Title: pw_user.c « pw « usr.sbin - src - FreeBSD source tree

that's the evil I was speaking of

I don't understand why it's evil.

There are good reasons for it, outlined fairly cogently.

there is one time where you want /home -> /usr/home, and that's when you're using ufs, multiple partitions, and no dedicated /home partition

which, these days, is approximately nobody

as long as you create /home yourself first, that code isn't used

if you have actual human users, then you almost never want /home to be part of either / or /usr

On ZFS, it's zpool/usr/home because of boot environments - and it gets mounted as /usr/home/

the only reason given in that comment is to put it in /usr rather than / if it comes down to a choice, since /usr is usually bigger if they are separate partitions

s/it\'s/the dataset/

well that's just wrong

debdrup: i dunno if my MUA allows that, but i can try

meena: mutt and derivatives lets you use -f /path/to/mbox to open arbitrary mbox files

RhodiumToad: I don't think boot environments work without doing it like that.

debdrup: and apparently, i could totally use it with protonmail… brian-thompson.medium.com/setting-u…-client-with-protonmail-49c042486b3

Title: Setting up the Mutt mail client with Protonmail | by Brian Thompson | Medium

there seems no reason why boot environments should care about /home any more than they care about any other random extra dataset that you might have

meena: neomutt has a bunch of rather nice additions

Hello, I switch recently form Mutt to Neomutt. Seems Neomutt is fastter than Mutt and I agree, there is nice additions

debdrup: i've got thunderbird installed, and can open .eml files (fancy name for mbox). Would you reply-to-all or just to someone specific?

meena: Just reply to the committer

not also to the list?

but then the question is which list…

Well, I typically do ^l which replies to a list if List-Post is defined in the headers..

meena: dev-commits-ports-main@

ate

aye, even

(i am hungry tho)

cool. I've got Thunderbird setup now to use my freebsd@ account

hrm… do I have to be subscribed to the list?

I'm pretty sure you have to be subscribed to every list to reply to it, but I'd recommend using the nomail unless you're following a particular tree closely (ie. using it for development).

s/nomail/nomail option/

debdrup: ports, not so much… src, yes, but i use FreshBSD as my eye on that

Yeah, I use freshports for src and ports, and am subscribed to the doc mailing list.

Err, freshbsd for src and freshports for ports (via an RSS feed)

freshports.org/backend/news.php?flavor=new only lists new ports.

Title: FreshPorts newsFreshPorts newsnews/py-sabctools - 7.0.1misc/py-torchvision - 0.15.2.r2misc/py-pytorch - 2.0.0devel/rubygem-rubocop-factory_bot - 2.22.0math/py-lrcalc - 2.1textproc/py-jupyter_sphinx - 0.3.2devel/py-pydantic-core - 0.28.0lang/scratch - 1.4.0.7x11/habak - 0.2.5graphics/py-colorthief - 0.2.1graphics/py-haishoku - 1.1.8devel/R-cran-timechange - 0.2.0math/octave-forge-octave-pool - 1.0.0emulators/qemu7 - (1 more message)

debdrup: what about deprecated / deleted ports

meena: *shrug*

I'd suggest talking to Dan about it.

He's usually very receptive to feedback :)

lists.freebsd.org/archives/dev-comm…its-ports-main/2023-May/060041.html :( Thunderbird breaks threading :(

Title: Re: git: 2aedfff8a362 - main - sysutils/tuptime: Fix cron file

then you are using it wrong (somehow), it doesn't for me

did you reply to the actual message in the list?

yuripv: i downloaded the message, and then hit: reply-to-list

so it's missing the message ids it can use, not a thunderbird fault

yuripv: lists.freebsd.org/archives/dev-commits-ports-main/2023-May/060018.txt this seems to have everything needed

Message-Id: <202305080736.3487asix068025⊙gfo>

i also don't see user-agent in your reply, your smtp relay stripping too much?

you could use test⊙Fo mailing list to check

do i have to subscribe to that one as well?

no idea, just try, you can spam it however you want :D

meena: the in-reply-to header item is what controls threading in mlmmj I think

also yeah, it might also be the smtp

Title: Responding correctly to threads in Thunderbird

That looks like the thing you want.

jwz.org/doc/threading.html this is the original source for the standard

Title: message threading

..oh, that's mentioned in the article, too.

Title: In-Reply-To header being removed if no References header is present · Issue #122 · ProtonMail/proton-bridge · GitHub

hi, I am reading about the Ghostscript API, it says you need to include the header (and likely the library).

the header is gsapi.h, any tip on where/how to search which packate does have this header?

gsapi.h? for what version? I see an iapi.h

/usr/local/include/ghostscript/iapi.h was installed by package ghostscript9-agpl-base-9.56.1_10

more generally, the pkg-provides package allows you to look up what packages contain any given file

Thanks RhodiumToad , are you suggesting that gsapi.h is actually iapi.h?

gsapi is Kerberos stuff

gssapi is kerberos-related

As far as I am reading in ChatGPT (could be totally wrong), iapi.h is the low level, and gsapi.h is the actual API to use when embbedding Ghostscript in an application

I don't seem to have a gsapi.h

... why are you reading anything from chatgpt?

Sometimes it says something

I am reading iapi.h and it seems it has the functions I am looking for

I guess it is time to read the API documentation at Ghostscript

Thanks RhodiumToad

13:54 <RhodiumToad> ... why are you reading anything from chatgpt? ⬅️ why indeed…

hey folks. turns out i had dtrace enabled

so if you want Rhodiumtoad i can do some further poking

what is the preferred fw for Freebsd? ipfw?

the one with the features you want

There's still ipf.

I'm not sure why apache24 has started quitting on its own. It has happened three times since I upgraded to 13.2. It never happened before (back to 10.x).

Schamschula, Is it on the public Internet where it gets network abuse from external influence? Anything in the error log? Could it be running out of virtual memory? I monitor Apache and restart it with "monit".

rwp: Publicly accessible, but content is not exposed. Unfortunately, the error log has rolled over with the last restart. I'll have to change the configuration to keep a few older logs.

if it's crashing, there should be a message in system log

kernel: pid 14922 (httpd), jid 0, uid 0: exited on signal 4 (core dumped)

Schamschula: and, is there a core?

Mmm, core files.

Looking for them

..although they won't do much good without debugging symbols.

Title: mpm_common - Apache HTTP Server Version 2.4

I seen nothing in /var/crash

that's where kernel crashsumps end up in

demons are a bit tricky, because they usually cd to /

but nothing can (hopefully) write to / other than root, and apache just starts as root, it doesn't do any real work as root

Apparently, I have to set kern.sugid_coredump to 1. Unfortunately, the module documentation is unclear on how to do that.

sysctl(8)

man.freebsd.org/core(5) has info on how to generally collect coredumps, which isn't Apache Httpd specific

Title: core(5)

I set `sysctl kern.sugid_coredump=1` Hopefully, next time it happens there'll be bit to go on.

You might want to set the value in sysctl.conf(5), if you want it to persist across reboots.

debdrup: done

Schamschula: also, like I said, you want debugging symbols - which means you need to build apache (and possibly everything that hooks into it) with debugging symbols. Check wiki.freebsd.org/DebuggingPorts

Title: DebuggingPorts - FreeBSD Wiki

Hello, when setting up an NFS server, instead of listing clients by individual IP can I use CIDR notation to allow anyone in the range of IP addresses to connect to the NFS server?

yuripv: strange coincidence, I was looking at enable_altgr just yesterday. I wondered why it didn't work while enabled by default. From your commit it seems it is off by default, is that right?

complexnumber: That's covered in exports(5).

thank you, will look it up now

freshbsd.org/freebsd/src/commit/7e0f8b79b77 woo!

Title: FreeBSD / src / 7e0f8b7 / Add mrsasutil(8) as alias to mfiutil(8) - FreshBSD

Schamschula, Also look at the apache access.log to see what was happening right up to when it crashed. Can be a clue there. Also if using the PHP module it might be PHP crashing instead of Apache.

pstef: no, it's enabled by default, i was looking at the forums post asking how to make right alt behave like left one

rwp: found the culprit: my log rotation script. Now, I need to figure out why that suddenly no longer works correctly.

It looks like using apachectl to do a graceful restart might not be the best way to restart httpd. We'll see if using service apache24 will fix this.

Schamschula, Yay for finding the root of the problem. But "apachectl graceful" is the normal thing to do. If that is causing a crash then that's the bug source.

ScentedFern, log rotate - kill -30

30 is USR1

Do you have crash apache24 during logrotate?

If yes and you are using mod_php, then check this: bugs.freebsd.org/bugzilla/show_bug.cgi?id=268318

Title: 268318 – www/apache24 with www/mod_php82: opcache + ASLR turned on crashes Apache

Schamschula

yuripv: ah. I thought after your recent commit the page doesn't say but seems to suggest that the sysctl defaults to 0

Schamschula: Apache ships with a rotatelogs command you can have it pipe to, saves having to kick Apache every cycle

pstef: i was looking at enable_bell, it's 0 by default and page says =1, so i used the inverse logic :)

hmm, maybe you're right

but the kbd_* ones show the default (i'd just remove the =N part from synopsis)

altgr=1 doesn't seem to work for me in VT anyway, but I'll leave that one for some other time

pstef: with which keymap?

I just added a hack to clear color inversion of selected text when screen changes and diff is non-zero, but I haven't checked if that works :)

no idea about the keymap, I wasn't even aware

in some keymaps altgr acts e.g. as layout switch

just tried the pl.kbd, altgr seems to work there