Further Planning On Ubuntu's New Package System

[phoronix]

Phoronix: Further Planning On Ubuntu's New Package System

Yesterday during the virtual Ubuntu Developer Summit to begin working out Ubuntu 13.10 plans were more discussions surrounding the distribution's proposed new packaging system...

http://www.phoronix.com/vr.php?view=MTM3MjU

[uid313]

Bah

I rather see better support for PackageKit and AppStream.

[madjr]

Hopefully they can use listaller or 0install and not do like gnome did and create another one (glick2)

http://blog.tenstral.net/2013/02/listaller-appstream-gnome-apps.html

[Teho]

Hopefully they can use listaller or 0install and not do like gnome did and create another one (glick2)

I'm pretty sure this is entirely new system.

The Ubuntu developers have also looked into similar existing tools such as Listaller or 0install but there are some things which they prefer to do differently; e.g.: Listaller is dependency-based and they prefer this to be as independent as possible and 0install would also need some system integration problems to be solved, so instead, they've decided to create a new installer.

-webupd8

[benjamin545]

lol. ubuntu has a bit of a trend

if current supported and accepted ways of doinging things exist, and the only obsticles are

some system integration problems to be solved

just compleatly overlook it and run its name through the mud andreimplement your own hacked together implementation that nobody else in the community cares to see exist and then tout it as a revolutionary and game changing awsomesauce new thing that is going to ROFL stomp all the others.

why the hell did ubuntu ever become the defacto newbie distro? why couldn't we just leave it as a debian testing + nvidia blob and flash on a braindead install disk? why didn't fedora take the intro distro spot?

[staalmannen]

I for one actually like this idea. In a way it is similar to what windows programs often do: bundle critical dll:s with the application. One great advantage with this is that a packaged application will have a longer "shelf-life" and will be installable and runnable on several releases of Ubuntu. It should also be fairly portable to other Linux distros (I have tried some repackaging of debs and rpms on Arch and got very frustrated with some (in Arch terms) old library dependencies expected by the binary) What one can hope for is that the "portable apps"-type packages installed via Click will have a custom LD_LIBRARY_PATH where the Click package library directory is put last (can we have private namespaces on Linux?), so that more up-to-date and security fixed .so:s are used first and that bundled libraries are only used when the system does not provide the dependencies.

[strcat]

I for one actually like this idea. In a way it is similar to what windows programs often do: bundle critical dll:s with the application. One great advantage with this is that a packaged application will have a longer "shelf-life" and will be installable and runnable on several releases of Ubuntu. It should also be fairly portable to other Linux distros (I have tried some repackaging of debs and rpms on Arch and got very frustrated with some (in Arch terms) old library dependencies expected by the binary) What one can hope for is that the "portable apps"-type packages installed via Click will have a custom LD_LIBRARY_PATH where the Click package library directory is put last (can we have private namespaces on Linux?), so that more up-to-date and security fixed .so:s are used first and that bundled libraries are only used when the system does not provide the dependencies.

They'll just depend on what Canonical defines as the "base system" and then statically link in any other dependencies, base on the Canonical blog posts about this.

Portable binaries can already just link against libc and statically link the other dependencies, and there's the Linux Standard Base defining a set of libraries that can be dynamically linked against.

[ImNtReal]

Why re-invent .msi files? There are reasons I'm trying to get away from Windows, and the fact that the traditional package systems used by most linux distros is superior to the crap they've been doing in Windows is a big one. This is a major step in the wrong direction.

[uid313]

Why re-invent .msi files? There are reasons I'm trying to get away from Windows, and the fact that the traditional package systems used by most linux distros is superior to the crap they've been doing in Windows is a big one. This is a major step in the wrong direction.

Well at least MSI is 1000 times better than InstallShield and Install WISE.

Also with MSI you can deploy them on Active Directory which is pretty cool.