SoylentNews is people
SoylentNews
For years, Google used an in-house Linux distribution called Goobuntu (based on Ubuntu LTS releases), as its development platform. No more.
After more than five years with Ubuntu, Google is replacing Goobuntu with gLinux, a Linux distribution based on Debian Testing.
[...] As MuyLinux reports, gLinux is being built from the source code of the packages and Google introduces its own changes to it. The changes will also be contributed to the upstream.
[...] How does Google plan to move to Debian Testing? The current Debian Testing release is upcoming Debian 10 Buster. Google has developed an internal tool to migrate the existing systems from Ubuntu 14.04 LTS to Debian 10 Buster. Project leader Margarita claimed in the Debconf talk that tool was tested to be working fine.
Google also plans to send the changes to Debian Upstream and hence contributing to its development.
[...] Back in 2012, Canonical had clarified that Google is not their largest business desktop customer. However, it is safe to say that Google was a big customer for them. As Google prepares to switch to Debian, this will surely result in revenue loss for Canonical.
(Score: 0) by Anonymous Coward on Tuesday January 23 2018, @10:47PM (1 child)
So you are saying thatg systemd solves the right set of problems in the wrong way. ? As a Arch linux user I find that the systemd guy was the only one to come with a comprehensive solution and I understand the distro packager choice of going with systemd.
(Score: 1) by pTamok on Wednesday January 24 2018, @08:27AM
systemd is excellent at solving certain problems found with other inits. The problem is, it doesn't play nicely if you want to mix and match modules/functions with other approaches, and it is a bit big for people with embedded and/or small systems. As it has assumed more and more capabilities, it has become a dependency for many other applications whose programmers, naturally, want to do the least work for the most benefit, and thus only offer systemd compatibility, meaning that anyone who wishes to eschew systemd suddenly has the need to fork large amounts of code, or write a separate shim - likely in an area they do not specialise in.
We are at the de facto point that we no longer have GNU/Linux for most people - we have GNU/systemd/Linux. There are some small distributions that don't use systemd (without-systemd.org gives a list [without-systemd.org] ), but 'mainstream' distributions are all systemd users. This is close to a monoculture that some people regard as dangerous, for many reasons. As systemd doesn't suppot non-Linux kernels, maybe it will give more impetus to GNU/Hurd (he says, drily) - or, as we have seen, a lot of people have moved to one of the BSDs. For people who regard the GPL as important, this latter is not regarded as a good step, and tends to be rather contentious for those who care about such things.
I expect in a couple of decades, something else will come along to be concerned about. In the meantime there will be a lot of bitter infighting about things that most people neither know or care about. Whether they should is another debate.