Re: PKGBASE Removes FreeBSD Base System Feature

From: Sulev-Madis Silber <freebsd-current-freebsd-org111_at_ketas.si.pri.ee>
Date: Sat, 09 Aug 2025 02:03:00 UTC
i think vermadden actually "screamed" because this turned into kind of a "desbsd" discussion. isn't "vmdbsd" either. is, well, "free" bsd, funnily

before this turns into complete virtual fistfight...

who has authority on this os anyway?

i already once mentioned elsewhere that not much goes on in fbsd. as compared to eg, ff. i consider mozilla having gone past "user is our friend approach" lately. unfortunately without users, projects become meaningless

but here?

who has, like, powers? is it:

old devs who have write access to src repo?
new devs who have write access to src repo?
currently elected core?
any of previous core?
any original guys like mckusick?
big companies?
small companies?
sysadmins?
users?
random hackers?
sunday fbsd users?
the foundation?
any other known guys?
or i mean gals, we have those too, tho i only recall one once had fights over inclusion problems which i don't mind of who wrote the code
any nation state?
who else?

at which level any of those should affect things?

how much time (like code or so), or money does one need to invest before (s)he has voice? how much value (s)he needs to obtain from fbsd before (s)he has, again, "voice" / permission to be heard...

because it affects direction of fbsd goes to

when there was pkg switch, i found it too fast and without much real reason. somebodys fs got full i-node wise? old thing had more files. new one has sql(ite) db. there was also concern on what if it breaks. now there are 7 backups of it, made daily

pkg however did improve things

but switch was for unknown reasons, funnily. i didn't have pressing need as sysadmin, so who did?

who makes decisions for me and why?

decisions, when made without you in loop can be either good or bad

maybe we need like freebsd user committee too? but then, what happens to users who don't want to be in it?

i haven't seen those discussions yet much. maybe they have been hidden in number of different places. there's literally a ton of places where people talk about fbsd. officially or not. it's not possible to follow them all. basically there's some kind of fbsd discussion on pretty much everywhere and in every language. because apparently it gets used. which is a good thing

but yeah, on pkg, on ports, somehow poudriere is a thing. that's also weird, because who says i can't build ports manually?

i recally this linux user angry comment that fbsd only has two fses. a ufs and zfs. a bicycle and a rocket. nothing in between

it's kind of same with other things. to pkg it was upgrade from bicycle to rocket. pkgbase is same. so is poudriere. which has this horrible hard name that in french means a building where gunpowder is stored. unsure why this name. because all the power is there? or that they sometimes blew up, leveling better part of a town (part)?

poudriere was like addon but now it's like standard? ports don't fully work with portupgrade, etc. and it's a hard thing to use. seems like it's best use case is in fbsd build cluster. funnily this is where it's used and this is for what it was built for. elsewhere it's usage is complicated

anyway, a rocket is good, gets you straight to the moon. it's also a "rocket science". what happens if you don't need it? where's car and bus and taxi and motorcycle and truck and so on here?

learning to fly a "rocket" is awesome skill. but it's a huge learning curve. well it depends. can you use vi? vi is a editor that's often used as a example for thing that for legacy reasons has been something that's extremely hard for new user to use. i still have vi usage issues despite decades

this rocket flying skills is why we have such long discussion here anyway. many, me too, find rocket really good but how to fly? like, it's hard. and people don't like "hard". people get angry when hard. i get angry when hard. for a number of times in my life i have smashed a thing that angers me

fbsd was very different when i came from linux. now it's like good, somehow. i have wondered where to go from here if needed

like, where to go if 32bit is gone? old fbsd? i mean not really i386 but eg arm. reason was good too. no dev power. the problem is that 32bit arms are still selling. unsure who sells pure i386 but arm is here. i also have vague reasons why 32bit gets into way of developing stuff. but as user, it's different. hw is sold. hw is cheap. but fbsd doesn't support it? i hope fbsd actually supports it until hw is not sold anymore and most of it is physically dead or in museum. mips is out already but i swear i saw mips being sold still

it's also different to look it from user side. i know arm is fragmented but user just wants to run it

same with pkgbase maybe

a disconnect between user and dev

i guess this is endless discussion, so i could just ask who has power to decide?

when i came to fbsd this was like oh a slow pace. now it's kind of too fast. i can already see linux failures. a rush is bad. rush gives you everything. at a cost

unsure if ixsystems stops funding fbsd if they don't implement pkgbase fast or what? quick look showed that company worker being first, 6 years ago, asking people to test it. maybe it had a earlier development somewhere else too

is that the needle that pricks into somebody's ass?

because i have like moderatively low interest in pkgbase. all i see is, now it's easier to include stuff for embedded usage. and now every file is tracked

rest is harder. the *.pkgsave. the skills maybe. unsure, i'd have to try. and on different systems and env

i think indeed pkgbase goes in. it might get some updates to plug the screams

btw, i never understand the shoot first and ask questions later approach

i still have issues with libxo too. idea is good, but it's a rocket. libxo also came from, was it, juniper? i once moaned about it. feedback was mixed, from are you trolling to ok maybe we need to take second look to it

so does money control fbsd after all?

also, go and look how dragonflybsd was created. make sure to check into mailing lists to see people insulting each other. seems bad

btw i like the fbsd approach on how you could build stuff locally. before it was even too much maybe. now it's too less? there's a cost here. there's more control in local builds. time, energy. also less external attacks. i don't know what linux distro users do. do they always trust maker of that distro? only focus on ease to install and use? linux users also told me that fbsd ports look bad because as if every port has like personal owner to it. despite linux packages have people behind them too. i would be they also have different people working on stuff. it's just hidden from users who don't see it. "system" does it for them. some people have told that now they can like finally use fbsd because they can just install stuff with sane package manager and don't need to wait. i mean there was pkg_add before you know. worked, even if poorly

i'm unsure, maybe that's why pkg came to be? i personally don't have issues with local builds and if you are large fbsd user (i'm not right now but i imagine how others do it), it's probably even good idea. less load on projects volunteered machines and more secure. and maybe not much time and power either. esp with latest hw

so yeah i have ton of questions about every design decision of fbsd. where it, any one thing, came. why. benefits? drawbacks? mitigation? etc...