FreeBSD development seems lost

By now I think most of the fuss about pkgbase is a storm in a teacup.
I disagree looking forward. They haven't started "picking" at the base packages yet. The firewalls, nvi, csh, etc will almost certainly be pulled out of base within half a decade. Perhaps zsh will be added instead because "thats cool currently".

You are basically going to be left with a teacup, but the storm inside is going to be replaced by a scoop of random soil, sand and dust from the back garden. Each time you want to fill the teacup, its going to be a different mix.

I would have more confidence if I had seen any Linux distro with this package mechanism with a consistent base. But I haven't. Not even Alpine with 10 default packages.

My experience with freebsd-update is that it is indeterministic & lame, takes hours and one can end with a broken system.
I upgraded debian many times, really good experience, works without troubles and is not an endless process.
Debian's install is completely random. Their minimal, standard package selections change by around 2% of their packages every release. Its not really about being a pleasant experience, its about what you are left with at the end.
 
From one retro enthusiast to another, I think dropping i386 is a crime.
i386 needs to go, and be replaced by 32bit ARM, and Risc. Also, let NetBSD have i386 for legacy purposes. 32bit, needs to be for simpler operating systems like Haiku, RiscOS, PDOS etc... ARM and Risc are better for 32bit computing, than i386. I read that i386 has kept backward compatibility for 286, 386 and 486 which made it not as efficient. So ARM and Risc could run on smaller hardware.

And to that end, I'm certain that MBR will be on the chopping block soon.
I wish there were an MBR version 2, which had 8 or 16 partitions, instead of 4. I don't like the idea of how upgradeable UEFI is. MBR was set into the hardware.


I think that the rust-in-base
I also wish claims of Rust in base would stop being brought up. Rust has its place, and they settled it, that it's not going into base. FreeBSD does need a C compatible safer language for base, but I believe it's not Rust. Also, there's nothing wrong with Rust in ports. If you want Rust in base, use RedoxOS, and use C for the userland.
 
It's almost like end users shouldn't have any negative opinions. I get it in the sense that negativity doesn't breed anything positive or productive, however, when your system works great under 14.1, suddenly things break and/or are buggy in 14.3 and 15.0 arbitrarily drops support for your peripherals; it pisses you off.

And while I don't agree with every gripe, I wholeheartedly agree with the title of this thread. QA practices need to be improved. Introducing bugs and breaking basic installer functionality in a minor release is unacceptable in my opinion.
 
I disagree looking forward. They haven't started "picking" at the base packages yet. The firewalls, nvi, csh, etc will almost certainly be pulled out of base within half a decade. Perhaps zsh will be added instead because "thats cool currently".
As long as it is that pkgbase packages are build from /use/src it is only a question of how that is sliced. Pulling in stuff from /use/ports into base, now that is what Linux is doing and getting into hot water with. And I have seen enough Linux systems being restored from snapshots after an upgrade, that is not what I would want to see here.
 
Actually FFS is solid as hell. It is slower though but for my use-case (I am not an enterprise), it is easily good enough.

vmm and chroot are also good enough for me. I only allocate one core to my VMs anyway and my use of a jail is mainly building software / administrative rather than security, so chroot will suffice.

For me personally, OpenBSD is an easy compromise.
Thank you for the feedback, it's an interesting point of view.
I might dig a bit on this, not that I have a desire to leave FreeBSD but just because I like how this project handle things, small team with a specific vision.
It's good to look at what others can do/offer, it gives ideas and overall it makes us progress in the right direction, I think.
 
Strange. "Hours" I saw only one time -- when ZFS ARC was too small.
Usually f-u works fast.
Perhaps that is the reason. I have 'only' 4GB in my Desktop. Perhaps I upgrade to 8 GB.

The other solution is to install FreeBSD with UFS if I decide to work with FreeBSD 15 and when I do it.

For A NAS I will need ZFS, or perhaps Hammer 2 ?
 
Someone else said they removed Perl from the base recently, I must admit I was a bit surprised... I guess sed and awk are still there!
"someone" was wrong. perl was removed over a decade ago. It became too complicated to try to keep perl-in-base up to date w/rt our release schedules.
 
> I will not be surprised if 16.0 adds some sort of Flatpak-type system, or Wayland, or a systemd-style init nightmare.

This is simply FUD. All you are doing is demotivating people who are doing the heavy lifting.

For myself, I can either spend time (like this) trying to correct misperceptions, or I can go work on problems, or I can go sit out on the back porch and play match-3 games. The latter two are far less stressful than the first.

Please keep this in mind when posting. Thanks.
 
Tragic.

Perhaps a metaport "classicbsd" with all that programs and the games will be necessary.

Another called "mitxclients" is necessary already now.
At least nvi CANNOT be pulled out, as it is installed as vi, which is part of POSIX.

To pull it out, POSIX needs to pull out first.

So as ex, which is actually the same binary as vi (hardlinked, historically) and works differently with argv[0], which is the command name invoked as.
 
Completely unnecessary change for luanatic reasons.
(with respect to forth)

It had gotten to the point where very few developers wanted to work with forth. There was indeed discussion about including lua in the base system, but the compromise was to restrict that minimal version lua to base system only, thus avoiding the constant problems with perl being stale.
 
It's almost like end users shouldn't have any negative opinions.
When the only thing the _volunteers_ get is negative feedback, it is completely demotivating.

For paid employees, it's different.

I agree that regressions are bad and I spend some of my own free time trying to highlight them in Bugzilla. But I am just one contributor.
 
So we have 6 pages of fear and loathing by now and the only real thing that is reflected in code that was brought up is pkgbase.
I was reading this while watching TV and eating a sandwich, and even so, it was a complete waste of my time. (cracauer's and cy's as well.)

There is a reason why I do not often visit the forums, and this thread is an excellent example.

Sigh.
 
I was reading this while watching TV and eating a sandwich, and even so, it was a complete waste of my time. (cracauer's and cy's as well.)

There is a reason why I do not often visit the forums, and this thread is an excellent example.

Sigh.
Sorry to hear that... sometimes uninformed users do stir up shit just because they like to talk. Bashing something used to be 'cool', but persistent bashing by the uninformed masses who don't contribtute otherwise - that, IMHO, is going too far, esp. when it snowballs.

I'd suggest looking at this as 'storm in a teacup'.

Frankly, I'm grateful that this is FreeBSD - in the Linux camp, I've seen much more colorful and poisonous comments, which are also quite emotionally charged...
 
Back
Top