Debian 10.0 "Buster" Release Images Are Up For Testing

[Brisse]

Seems like a really good release to me, but as a sid user I'm just glad the freeze will be over soon so that I can go back to the convenience of not having to pull in stuff from experimental and build kernel from source for a while

Buster will have a place on one of my laptops though. Don't use it as often as my desktop so I don't appreciate the constant updating of a rolling distro there.

[Ipkh]

I hadn't tried debian in years. I could never get Sarge to install properly on my system. Even buster rc2 has a few issues. It wouldn't update my grub tables on 1 system, but instead perfectly on another sans non-free firmware so no internet. Overall it's the most trouble free release of debian ever. It even loaded up a desktop on my nvidia cards once the drivers were installed. No other debian version I tried managed that.

[waxhead]

I hadn't tried debian in years. I could never get Sarge to install properly on my system. Even buster rc2 has a few issues. It wouldn't update my grub tables on 1 system, but instead perfectly on another sans non-free firmware so no internet. Overall it's the most trouble free release of debian ever. It even loaded up a desktop on my nvidia cards once the drivers were installed. No other debian version I tried managed that.

Debian has come a long way since Sarge. These days everything does work out of the box - and in case it doesn't you get the images that has the non-free stuff enabled by default and off you go! I personally use Debian for everything as it covers all I need.

[Sonadow]

Debian has come a long way since Sarge. These days everything does work out of the box - and in case it doesn't you get the images that has the non-free stuff enabled by default and off you go! I personally use Debian for everything as it covers all I need.

Debian's package manager still sucks balls and pisses me off.

When I want to install a large framework, I want to do it with just running <package manager executable> install <package name> Like, you know, dnf install large-framework.

Not like Debian and Debian-based distributions where one needs to run apt-get install large-framework-subcomponent1, large-framework-subcomponent2, etc etc... only to find out much later after multiple errors that large-framework-subcomponent7, large-framework-subcomponent32 and large-framework-subcomponent48 were omitted because the Debian maintainers thought 'they weren't critical to large-framework's operation'.

[Ipkh]

Easy fix. Install the synaptic package manager.

[GreenReaper]

When I want to install a large framework, I want to do it with just running &lt;package manager executable&gt; install &lt;package name&gt; Like, you know, dnf install large-framework.

Not like Debian and Debian-based distributions where one needs to run apt-get install large-framework-subcomponent1, large-framework-subcomponent2, etc etc... only to find out much later after multiple errors that large-framework-subcomponent7, large-framework-subcomponent32 and large-framework-subcomponent48 were omitted because the Debian maintainers thought 'they weren't critical to large-framework's operation'.

It sounds like you might want:

APT::Install-Suggests "1";

in your /etc/apt/apt.conf (or a file in /etc/apt/apt conf.d/).

Conversely, I run servers, so I leave suggests at the default and have Install-Recommends set to 0.

Personally I have the reverse problem: I want PostgreSQL to be installable without LLVM, which I see as an unreasonably large addition (and which more than doubles the size of the package itself with files to support it).

It's a matter of preferences, and Debian makes it easy for end-users and package managers to express and use those preferences.

Oh, and a note with PostgreSQL: if you're upgrading to buster, be sure to take the DB down and reindex after the upgrade before you put it up again, as glibc changes relating to collation mean you could end up with some nasty data corruption otherwise (this is not specifically a Debian issue, it applies for any use of later libc). We ended up having to remove duplicate primary keys, which was fun when you had to do that in a way that avoided the indexes.