I used to end up with some real Frankenstein systems so I try not to screw with the ‘recommended’ versions. Kicad is one of two exceptions now. The other is SageMath. Both are end user apps anyhow.
Just checked and it looks like Debian finally supports SageMath. It wasn’t even an option in Jessie. It used to be an easy compile for Debian and the developers used it but just gave up trying to jump through the Debian hoops. At the time I think their live distro even shipped on Debian.
There are no official PPAs from the KiCad Team, so far I know Jean-Samuel isn’t a official member, he is providing packages through PPA on his own and users of Ubuntu appreciate this. But that’s not the point. Ubuntu has no active maintainer for KiCad packages but they pull in the packages from Debian. So long nobody is take care on the packages in Ubuntu for older LTS versions there wont be anything newer than the version the LTS release was made. PPAs are just a way to workaround this circumstance. So that’s what people think why Ubuntu is “better” than Debian. Canonical is a company and they want to earn money so they making the rules.
I will prepare backports of KiCad for Buster and Stretch as long it is possible. I do this in my spare time as also Jean-Samual is doing.
And where is the difference to an PPA?
If users don’t know about this they also do have to stick around some older versions, not a real good reason. And backports are mentioned on the download site for Debian as well as PPAs for Ubuntu.
That’s a thing you can’t blame Debian on and by the way, the same problem is / was happen in Ubuntu. The reason here is a technical reason related on the various STL standards of C++. And I did have to deal with that already for version 5.1.0 simply because downgrading a packaging isn’t the correct way and not possible on the build daemons. So please distinguish always correctly.
There’s discussion going on about skipping 5.1.3 and moving directly to 5.1.4 due to a bug which was introduced after 5.1.2 but before 5.1.3. Nothing dangerous. Packages and installers for 5.1.3 on many platforms are already out, it would just go unannounced and would be short lived.
I would definitely suggest a name change for the “stable” series of nightlies. The concept of nightlies is already alien enough to many inexperienced users let alone having two so vastly different things share the same name.
Maybe reducing the number of builds and calling them release candidates might be an option to get around that. (With reduced builds it might be even possible to expand the number of platforms available.)
The nightly 5.1 fork build would still be useful for testing purposes, I wonder if the 5.1.3 issue was due to the hiatus in the builds before the tagging, so a bug was not tested.
What would be great was if this daily build was just the bin section, no libraries, to get a small file size.
I’ll update the release announcement and there will be no official 5.1.3 release.
5.1.3 won’t be officially announced, 5.1.4 has been tagged and will be released and announced soon. There are 5.1.3 packages out there but it shouldn’t matter. This hassle was due to one bug which caused updating the footprints on a board to loose their positions. But now we got some other bugfixes, too, without waiting much longer
String changes were reverted in order to tag 5.1.4 using the same translations as 5.1.3. After tagging, the reversions were themselves reverted. The net effect was to get the bug fixes into 5.1.4 but keep the strings unchanged.