when lmde 2 (betsy) will be available?

[clem]

You're spot-on on many things here. Momentum amongst 3rd party developers is what really matters.

It would be futile to decide not to use systemd long term if no efforts existed (either from us or from upstream) to develop and maintain proper alternatives... and I don't mean just the init system here, but the whole range of services systemd might obsolete (upower, consolekit..etc).

I talked a lot with the guys who are splitting from Debian but there is no plan for us to switch base. There is no plan either to go against Debian's decision to use systemd. Right now in 2014 our only concern is to "choose" what to use for the 2+ years coming. Next cycle around, say in 2016, if we haven't done so already we'll likely align ourselves with upstream.

Of course, if the way upstream develops makes it impossible for us to ahieve our goals or to meet our own expectations, we'll get involved in the development (whether that means patching, forking, replacing with alternatives or writing our own). If and when that happens, we'll need to assess risks, costs, and goals. Talking about this is premature though. Right now systemd works pretty well, and the only issue on the horizon is that we're as stubborn as its development team when it comes to getting what we want (and it definitely looks like our priorities and philosophies differ)

On the matter at hand, we're only concerned (for now) about whether or not to use systemd in Betsy. At the heart of this there's power and session management in MDM, Cinnamon and MATE, and maintenance on top of Jessie. From a development point of view we're excited at the thought of jumping onto Systemd. From a maintainer point of view, I'm still afraid it's relatively new so I want to make sure everything works well before we consider using it.

One thing I didn't speak about also.. but that's premature as well... is the fact that systemd support in Cinnamon is activated at compiling time. I'm hoping we can solve this in 2.6 so that whatever choice we go for, people can decide otherwise for themselves and not have to recompile their DE to make things work.

[gweaver]

It would be futile to decide not to use systemd long term if no efforts existed (either from us or from upstream) to develop and maintain proper alternatives... and I don't mean just the init system here, but the whole range of services systemd might obsolete (upower, consolekit..etc).
..
Next cycle around, say in 2016, if we haven't done so already we'll likely align ourselves with upstream.
..
Of course, if the way upstream develops makes it impossible for us to ahieve our goals or to meet our own expectations, we'll get involved in the development (whether that means patching, forking, replacing with alternatives or writing our own).
..
Right now systemd works pretty well, and the only issue on the horizon is that we're as stubborn as its development team when it comes to getting what we want (and it definitely looks like our priorities and philosophies differ)
..

Thanks for enlightening us

I'm inclined to think that the least risky and least costly approach in the short term would be to use systemd in Betsy (I can't believe I'm saying this!), but to minimise any Minty dependencies upon it in order to keep your options open. If systemd does turn out to be the future, you won't have to jump init systems again. I'm pretty sure that sysvinit isn't the future.. No doubt someone will be along shortly with some good arguments for LMDE sticking with sysvinit!

Right now the momentum definitely seems to be with systemd, but longer term any of the following could happen:

• systemd might become rock solid and dependable, and the philosophy of the development team might change
• a sane systemd fork might reach maturity (uselessd ?!?)
• a simpler systemd compatible init system might be developed
• another compelling init system might reach maturity (openrc ?)

If I were a distribution developer I'd want to hedge my bets and be ready to jump on (or help with) whichever technical solution emerges as the most sane, elegant and mature. I guess that if a lot of distribution developers feel the same way another init system might win out.

Big decision time - hopefully most of the forum users seeing this thread will understand some of the difficulties the Mint team (and other distributions) face.

As an aside, this is an example of the quality package management for systemd on Debian Wheezy..

Code: Select all

root@test:~# apt-get dist-upgrade
Reading package lists... Done
Building dependency tree       
Reading state information... Done
Calculating upgrade... Done
The following NEW packages will be installed:
  libsystemd-daemon0 libsystemd-id128-0 libsystemd-journal0
0 upgraded, 3 newly installed, 0 to remove and 0 not upgraded.
Need to get 0 B/78.4 kB of archives.
After this operation, 242 kB of additional disk space will be used.
Do you want to continue [Y/n]? 
Selecting previously unselected package libsystemd-daemon0:amd64.
(Reading database ... 94366 files and directories currently installed.)
Unpacking libsystemd-daemon0:amd64 (from .../libsystemd-daemon0_44-11+deb7u4_amd64.deb) ...
Selecting previously unselected package libsystemd-id128-0:amd64.
Unpacking libsystemd-id128-0:amd64 (from .../libsystemd-id128-0_44-11+deb7u4_amd64.deb) ...
Selecting previously unselected package libsystemd-journal0:amd64.
Unpacking libsystemd-journal0:amd64 (from .../libsystemd-journal0_44-11+deb7u4_amd64.deb) ...
Setting up libsystemd-daemon0:amd64 (44-11+deb7u4) ...
Setting up libsystemd-id128-0:amd64 (44-11+deb7u4) ...
Setting up libsystemd-journal0:amd64 (44-11+deb7u4) ...
root@test:~# apt-get autoremove
Reading package lists... Done
Building dependency tree       
Reading state information... Done
The following packages will be REMOVED:
  libsystemd-daemon0 libsystemd-id128-0 libsystemd-journal0
0 upgraded, 0 newly installed, 3 to remove and 0 not upgraded.
After this operation, 242 kB disk space will be freed.
Do you want to continue [Y/n]? 
(Reading database ... 94380 files and directories currently installed.)
Removing libsystemd-daemon0:amd64 ...
Removing libsystemd-journal0:amd64 ...
Removing libsystemd-id128-0:amd64 ...

And repeat..

[py-thon]

apt-get autoremove acts stupid no matter what init system is installed. According to manpages "autoremove is used to remove packages that were automatically installed to satisfy dependencies for other packages and are now no longer needed." But you have to check very carefully because this "now longer needed" is wrong more often than not.

[Distro-Don]

apt-get autoremove acts stupid no matter what init system is installed. According to manpages "autoremove is used to remove packages that were automatically installed to satisfy dependencies for other packages and are now no longer needed." But you have to check very carefully because this "now longer needed" is wrong more often than not.

Instead of autoremove I have been using orphaner for years now and so far I have never had a problem.

[gweaver]

apt-get autoremove acts stupid no matter what init system is installed. According to manpages "autoremove is used to remove packages that were automatically installed to satisfy dependencies for other packages and are now no longer needed." But you have to check very carefully because this "now longer needed" is wrong more often than not.

Instead of autoremove I have been using orphaner for years now and so far I have never had a problem.

Thanks for the information. I tried orphaner, but it seems to do the same thing as apt-get autoremove in this case.
Something, somewhere in Debian packaging (or apt-get) isn't right..

[gweaver]

Install Debian Jessie, then you have MATE 1.8.1 or Cinnamon 2.2
https://www.debian.org/devel/debian-installer/

If you want you can later always add the Mint LMDE repo for some Mint tools

I installed Jessie Beta 2 server (tasksel OpenSSH, no desktop) in a virtual machine and added the Betsy repos.

Code: Select all

deb http://packages.linuxmint.com betsy main upstream import

I added some of the mint meta packages for Mate & XFCE, but not Cinnamon, then added a few more packages by trial and error until it was possible to log in (MDM wasn't happy). I didn't encounter any dependency issues. Bar MDM, it all works very nicely, but there isn't a lot to see yet. I haven't noticed any issues with either desktop.

Mate is at version 1.8 (as in Debian), with some good default configurations, but without (as of 12/12/14) the Compiz stuff that has made it into 17.1.
XFCE hasn't received very much customisation, but there are Mint metapackages for it so maybe it will become an official flavour?

Clearly there is still a lot of work to be done on Betsy, especially for Cinnamon, so it's probably best played with in a VM and not used on any machine that you rely on.
On the plus side Betsy/Jessie is more modern than LMDE and Mate is more polished than it is on Jessie. For me, being able to cherry pick packages from Debian Testing/Sid gives it a big advantage over Mint 17.x. I think it has the makings of an excellent distribution.

[py-thon]

I installed Jessie Beta 2 server (tasksel OpenSSH, no desktop) in a virtual machine and added the Betsy repos.[...]
I added some of the mint meta packages for Mate & XFCE, but not Cinnamon, then added a few more packages by trial and error until it was possible to log in (MDM wasn't happy).

That's a somewhat risky way. Mint's meta packages are meant for Mate 1.6 (LMDE's version) whereas the package manager will prefer (and did install) the newer ones from Testing (Mate 1.8 ).

[gweaver]

That's a somewhat risky way.

I like living on the edge.

Don't try this at home kids, this method could fry your computer, eat your files, or cause a plague of locusts! This is why I didn't post instructions. If you do try it, make sure you do it on someone else's computer! Actually, it wouldn't surprise me if the developers slipped summonthedevilhimself.deb into the Betsy repository, just to punish fools like me for sneaking a look at their upcoming distro before it's anywhere near ready.

Mint's meta packages are meant for Mate 1.6 (LMDE's version) whereas the package manager will prefer (and did install) the newer ones from Testing (Mate 1.8 ).

You are right of course. I didn't notice this because I started with XFCE, and added Mate during the trial and error phase. I haven't apt-pinned anything. It's surprising how little Mint is required to make Debian look Minty, and how well it all works. The artwork looks good already. I reckon the Mint developers could get away with using a lot of the Debian Mate packages and revving the version numbers of the few remaining Mint packages - if you hadn't seen Mint 17.1 you wouldn't notice the stuff that is missing (Compiz, language settings etc.).

In case you were wondering why I'd screw Betsy`s head onto Jessie`s body, there were two reasons:
1. To see if adding specific programming/analysis packages from Testing/Unstable causes any problems, prior to release (this will help me plan my projects and distribution upgrade path).
2. Curiosity - to watch Jessie morph into Betsy.

[filozofio]

I have installed Debian Jessie with MATE, and it works fine. I had to do several things to have a working MATE Desktop (for example, to use Dropbox, to have update notifications, and so on). The experience is good, and it really runs quicly, better than Ubuntu and Ubuntu based distributions. But I am anxiously wating for LMDE 2 Betsy. I hope that it will be a great distribution!

[Crewp]

Strange how Solydxk came out already using Debian 8, when it has not even been released yet, beyond an RC

[killer de bug]

Strange how Solydxk came out already using Debian 8, when it has not even been released yet, beyond an RC

Jessie is basically Debian Testing at the moment. Debian Testing is frozen since 3 months and thus pretty stable. That is why SolydXK decided to use it already.
This said, as they were based on Debian Testing, they were already using Jessie since some months!

[Crewp]

Thanks Killer de bug, I know Solydxk is like LMDE spin off since LMDE dropped xfce, kde, DE.I just thought they would wait until Debian 8 release, but as you say it;s been frozen now 3 months. Makes me wonder now if Betsy might be next.

[cb474]

Solydxk looks cool, too bad there's no Mate version. I guess that's what LMDE is for. I really will be happy when LMDE finally starts getting updated again. I've had to stop using it, because I just don't think it's secure anymore.

Anyway, here's the Solydxk post where they explalin their decision to go ahead and update to Jessie. http://solydxk.com/transition-to-debian-jessiewheezy/

I don't really get what their updating plan is now. Some places seem to say that it's a semi-rollling release (like LMDE used to be) but I can't find that language on their website.

[clem]

I just thought they would wait until Debian 8 release, but as you say it;s been frozen now 3 months. Makes me wonder now if Betsy might be next.

Jessie is getting pretty stable alright and we've been able to work a lot on Betsy. These past few weeks it's been my only focus. I'll post some news on Segfault soon to detail what's been done. You can also follow the roadmap to get an idea of what's left on our list: https://github.com/linuxmint/Roadmap/bl ... /README.md

As you can see there are no longer any critical issue on the list, they're all solved now and we're down to small problems, integration and/or cosmetic issues.

[Monsta]

Solydxk looks cool, too bad there's no Mate version. I guess that's what LMDE is for.

That's why SolydXK was born: http://solydxk.com/about/solydxk/

[Zill]

... I don't really get what their updating plan is now. Some places seem to say that it's a semi-rollling release (like LMDE used to be) but I can't find that language on their website.

AIUI, the "official" SolydXK (Home Editions) are now changing from a semi-rolling release (based on Debian Testing) to a fixed release (based on the new Debian Stable which will be known as Jessie). This "official" SolydXK will also replace the current SolydXK Business Editions which are already based on the current Debian Stable (known as Wheezy).

In addition to the normal Debian Stable security patches it is expected that, although based on Debian Stable, upgrades for a small number of packages (such as Firefox and Thunderbird) will be available. A backports repo is included to allow some packages to be upgraded on request.

In addition to these "official" versions, it is likely that SolydXK will still offer rolling releases to be known as "Enthusiast Editions" but such editions will need to be community supported.

See "Important SolydXK news"

[Crewp]

I just thought they would wait until Debian 8 release, but as you say it;s been frozen now 3 months. Makes me wonder now if Betsy might be next.

Jessie is getting pretty stable alright and we've been able to work a lot on Betsy. These past few weeks it's been my only focus. I'll post some news on Segfault soon to detail what's been done. You can also follow the roadmap to get an idea of what's left on our list: https://github.com/linuxmint/Roadmap/bl ... /README.md

As you can see there are no longer any critical issue on the list, they're all solved now and we're down to small problems, integration and/or cosmetic issues.

Thanks Clem, for the feedback, really looking forward to Betsy's arrival.

[cb474]

I just thought they would wait until Debian 8 release, but as you say it;s been frozen now 3 months. Makes me wonder now if Betsy might be next.

Jessie is getting pretty stable alright and we've been able to work a lot on Betsy. These past few weeks it's been my only focus. I'll post some news on Segfault soon to detail what's been done. You can also follow the roadmap to get an idea of what's left on our list: https://github.com/linuxmint/Roadmap/bl ... /README.md

As you can see there are no longer any critical issue on the list, they're all solved now and we're down to small problems, integration and/or cosmetic issues.

Thanks for the update, clem!

... I don't really get what their updating plan is now. Some places seem to say that it's a semi-rollling release (like LMDE used to be) but I can't find that language on their website.

AIUI, the "official" SolydXK (Home Editions) are now changing from a semi-rolling release (based on Debian Testing) to a fixed release (based on the new Debian Stable which will be known as Jessie). This "official" SolydXK will also replace the current SolydXK Business Editions which are already based on the current Debian Stable (known as Wheezy).

In addition to the normal Debian Stable security patches it is expected that, although based on Debian Stable, upgrades for a small number of packages (such as Firefox and Thunderbird) will be available. A backports repo is included to allow some packages to be upgraded on request.

In addition to these "official" versions, it is likely that SolydXK will still offer rolling releases to be known as "Enthusiast Editions" but such editions will need to be community supported.

See "Important SolydXK news"

Thanks for the explanation.

[Monsta]

Some LMDE 2 "Betsy" news: http://segfault.linuxmint.com/2015/02/about-betsy/

[killer de bug]

If this is finally possible:

For years new versions of software got into Linux Mint first and were then ported to LMDE. We’re thinking of changing that. New software would get into LMDE first and then into Linux Mint.

Then LMDE2 will rock!