What do you think about the new "package base"?

Personally, I think any person praising Debian should be banned on the spot.
I wouldn't go that far. IMVHO, Debian made a huge mistake making some p*ware an integral part of the system. But apart from that, I still think it's the most manageable of all GNU/Linux dists. If I, for some reason, need Linux, I'll take Debian (or maybe Devuan, for the p*ware reason).

Anyways, this thread still is about pkgbase. It might be a troll attempt, but I guess we can just ignore that and, well, discuss pkgbase :) I still think it's a good idea and hope for a sound outcome.
 
I wouldn't go that far. IMVHO, Debian made a huge mistake making some p*ware an integral part of the system.

Nah, I'm mostly hating multiarch and a myriad of little packaging tools with overlapping responsibilities. Building a customized package from the FreeBSD ports tree is infinitely more ergonomic and the best part is that I can actually remember the steps. In contrast I'm always confused which Debian tool does what.

Anyways, this thread still is about pkgbase. It might be a troll attempt, but I guess we can just ignore that and, well, discuss pkgbase :)

Except there is nothing to discuss, it's an utterly boring internal infrastructure change born out of desire to avoid maintenance of two similar tools: pkg and freebsd-update. I'm genuinely surprised this thread is not locked by this point.
 
Last edited:
I'd like to see a limited pkgbase, for compilers, toolchains, and for other rapidly changing developments like maybe newer video drivers. What if I can include only the video driver sets I want, like any combination of ATI, Intel, Nvidia or generic? It would also be nice to select firewalls, to include 1 or all 3 of them by choice. Just to keep the base system trim, but functional.

The install disc can be kept small, and we wouldn't need different installs for boot cd, full cd, and full dvd versions. Just 1 install cd for the running system of under 500mb.

Rapidly and constant developments have to wait for new releases, and that seems to slow it down.
 
I'd like to see a limited pkgbase, for compilers, toolchains, and for other rapidly changing developments

I agree but actually would also suggest the opposite would work too! A pkgbase for old trusty software. We are already much of the way there with using a specific package prefix. Basically Solaris 10 package systems on steroids. Imagine this kinda thing:

/usr/2010
/usr/2015
/usr/local
/usr/latest

This means that I could use Gnome 2 installed at /usr/2015/bin/gnome-session with Texlive in /usr/local/bin/pdflatex and an experimental clang with /usr/latest/bin/clang.

It would be a massive undertaking to maintain and far more than what the community could handle by itself but certainly if every other operating system disappeared and everyone focused on FreeBSD, it could be possible. Great for digital preservation too!

But yes, offtopic too ;)
 
I agree but actually would also suggest the opposite would work too! A pkgbase for old trusty software. We are already much of the way there with using a specific package prefix. Basically Solaris 10 package systems on steroids system.
...
It would be a massive undertaking to maintain and far more than what the community could handle by itself but certainly if every other operating system disappeared and everyone focused on FreeBSD, it could be possible. Great for digital preservation too!

But yes, offtopic too ;)
It's not offtopic, but it's too large a topic.
The scope of pkgbase should be limited. Gnome shouldn't be included, however, there can be allowed an offsite official Gnome and other hosted pkgbase repositories. Let them do that without interfering on the base, and while being compatible on top of the original pkgbase. Gnome doesn't suit everyone's idea of efficiency, but it would be great for them to have a repository. It fits everyone's purpose of varying opinions on what an OS should be.

Pkgbase should be limited to console, perhaps with the exception of a very limited minimal windowmanager, but nothing more. Other BSD distributions have small install mediums with a simplistic windowmanager included.

Pkgbase should be of high quality controls, and strict and limited software selections. Leave GPL3 out of it, and limit to GPL2 for older GCC choices when required for hardware. Let ports and offsite pkgbase repositories revolve around a lean, efficient and logical pkgbase and base system. By default, ports will become efficient, because many redundancies will already be satisfied by something basic. Packages/ports is for everything else.

FreeBSD installs have had microdistribution sets to choose from that were really from ports.
 
Last edited:
I misunderstood pkgbase from what I wrote above. Pkgbase is not the time saver I thought it would be. Everything still has to be built. But, it would save time for local mirror repositories.

What's not wanted is built anyway, at least for the first time. I don't see many benefits in that, unless src.conf is used as normal.

I want to trim time off of building the base, and not have to compile a compiler, then to compile or get a package for an updated one. I already trim the compiler out of base with src.conf, and be sure to have a compiler from packages.
 
Given the torrent of "What you think about" threads you started, all with a lot of negative nonsense and some Linux hate, I sincerely doubt that.

I personally don't hate Linux at all, I hate some particular decisions most distributions made, but that's even off-topic for this thread. The key difference between FreeBSD and Linux you (!) started to talk about is that FreeBSD is a whole operating system while Linux is just a kernel that's combined with a GNU userland by distributors. So far so good. Pkgbase won't change anything about that.
I could be ignorant but I'm not motivated to troll or rant. I do not hate Linux. My main OS is Linux. FreeBSD is just another option (2nd option).

I misunderstood pkgbase from what I wrote above. Pkgbase is not the time saver I thought it would be. Everything still has to be built. But, it would save time for local mirror repositories.

What's not wanted is built anyway, at least for the first time. I don't see many benefits in that, unless src.conf is used as normal.

I want to trim time off of building the base, and not have to compile a compiler, then to compile or get a package for an updated one. I already trim the compiler out of base with src.conf, and be sure to have a compiler from packages.
Many will hate me. But I really like the way of Debian. I like a bunch of .deb packages I could unpack and have a working system. I would like fbsd-base.deb to be the content of a clean and minimal installation with post installation script to set everything up for me. That way I could use debootstrap and don't have to go through many steps like we used to do to manual install fbsd from .txz from .iso media. That way I could upgrade everything as easily as apt full-upgrade. If you just copy the Linux way it's so good but it's not. Another meaningless/careless/unthoughtful move that solved nothing. Disappointed.

Base and ports should use the same compiler like on Linux. No need to bootstrap another compiler for the base. I don't care about such thing as developed as a whole or base and ports separation. Just pure worthless marketing statement.

I like to think in term of package not base nor ports. The package abstraction just like LEGO. Everything is just package and only package. With a right set of packages I could have a working system.
 
Not even that. Debian is still a good option in the GNU/Linux world. But this completely misses the point. We DO like the self-contained base system FreeBSD provides, for good reasons that don't need to be repeated here.


If you look at the very provocative OP in this thread, it seems completely pointless to me and shouldn't be considered any further.
 
I don't care about such thing as developed as a whole or base and ports separation. Just pure worthless marketing statement.
You can like or dislike this separation, but calling it a marketing statement shows that you are at best ill-informed.
 
Back
Top