nealmcb's questions

The update-alternatives command is a common way to specify which of various alternative programs should be the default in each of several categories of programs (web browsers, editors, etc). I think that the normal way for a user to change defaults is to run something like sudo update-alternatives --config gnome-www-browser though various apps rely on other ways of specifying default browser (see e.g. 'xdg-open').

But my question here focuses on meaning and policy around the priority configuration setting. The man page explains:

Each alternative has a priority associated with it. When a link group is in automatic mode, the alternatives pointed to by members of the group will be those which have the highest priority.

It also notes that Different packages providing the same file need to do so cooperatively.

I'd like to understand how these priorities are designed to work:

• Who determines the priorities?
• Are users ever expected to modify them?
• When packaging a program, how should these priorities be set?
• Is there an easy way to see the priorities that all the Ubuntu packages use for each alternative, as a guideline for current practice? I imagine that's easier for official packages than for those in add-on repos, but given a repo, I imagine that it could also be queried.
• What expectations should we have for organizations who package applications outside of the official repositories?
• Should packages ask whether the user wants to change the default upon install?

The question comes up for me after reading Microsoft Edge is Now Taking Over as Your Default Browser… on Linux, which documents how Edge, as packaged by Microsoft for Debian-based systems, is given a priority of 200 as a browser (with the comment # Good enough to be the default.). On my Ubuntu 20.04 system, the other priorities for both x-www-browser and gnome-www-browser are:

$ update-alternatives --display x-www-browser
x-www-browser - auto mode
  link best version is /usr/bin/microsoft-edge-stable
  link currently points to /usr/bin/microsoft-edge-stable
  link x-www-browser is /usr/bin/x-www-browser
/usr/bin/firefox - priority 40
/usr/bin/google-chrome-beta - priority 150
/usr/bin/microsoft-edge-stable - priority 200

I installed the official Ubuntu chromium-based browser chromium, which is now using snap. But it doesn't show up in the alternatives.

I haven't installed google-chrome-stable. Do they also use a priority of 200?

What happens when there is a tie in priority values among alternatives?

Many thanks.

TL;DR: How can I get myhost.local to resolve to my wifi interface rather than the Docker virtual network?

My desktop system's hostname is "myhost" (from /etc/hostname etc), running Ubuntu 18.04. The ip a command tells me that my wifi ip address is 192.168.0.6, and my rarely-used docker-network ip address is 172.17.0.1.

I like to use mdns/Avahi to access my desktop, via "myhost.local", e.g. when connecting to local jupyter notebooks from other hosts on my home LAN.

For some reason, mdns is assigning two addresses to myhost.local:

$ host myhost.local
myhost.local has address 192.168.0.6
myhost.local has address 172.17.0.1
myhost.local has IPv6 address fe80::46da:e48:c3c0:112

Jupyter notebook is running with ip=192.168.0.6

When I ping myhost.local, or connect to http://myhost.local:8888/ it worked a few months ago, but now it tries the 172.17.0.1 address, which doesn't work for jupyter.

This seems related to the unanswered question at the Docker Community Forum: Docker network interface interferes with mDNS address.

Update: to answer a question, this is the hosts file. I think it comes from the default way that Ubuntu sets up the hosts file, so localhost is not actually the same as myhost, but I might have tweaked it.

$ cat /etc/hosts
127.0.0.1       localhost
127.0.1.1       myhost

Bonus Update

user10489 correctly notes that other hosts will resolve myhost.local via mdns just fine. And my underlying problem was that I wanted to know what ip address to configure in my jupyter notebook to make it available over my wireless LAN. It's just that we don't see an easy way to use mdns on the local machine to get that answer.

The resolution to that underlying problem is this, to just figure out which ip address I have on a given interface (here illustrated with Ubuntu's awkwardly-named first wireless network name):

dev=wlp3s0
ip=$(ip -br -o a show $dev | (read interface status ip ip6; echo ${ip%/*}))
jupyter notebook --no-browser --ip=$ip

As part of the movement towards the Fediverse (federated Internet publishing and social media platforms) and away from the currently dominant platforms that exist to sell user data to advertisers, various organizations are providing their participants with supported instances of federated platforms like Mastodon, Pleroma, Peertube, and WriteFreely.

For example, Ubuntu provides members with email addresses and the PlanetUbuntu blogging instance as noted at Can anyone get an [email protected] email address? - Ask Ubuntu.

See also Debian's approach at https://wiki.debian.org/Teams/DebianSocial.

Does Ubuntu/Canonical plan to provide these sorts of collaborative tools, to facilitate participant collaboration, networking and outreach in a safe, secure and well-integrated environment?

Does PlanetUbuntu support ActivityPub for federated connections?

The Ubuntu vision of packaging apps in a way that automatically gets at least security updates still seems very important. So as I install Anki I again resist the advice to install it manually, directly downloading the latest from their web page.

Instead I want to see a snap package or a PPA which someone will keep up-to-date, but I can't find either.

There is an Anki package in Ubuntu Bionic, Version 2.1.0beta36, but I note that running it says

"This is an BETA build - please do not package it up for Linux distributions".

Am I missing anything, or are there plans to keep the Ubuntu package updated for at least security issues?

Note: I've also reported this to Anki: Packaged version for automatically getting updates: Snap? Ubuntu PPA? / Anki (Computer Version) / Discussion Area - Anki Support

The package pypy in bionic is compatible with Python 2.7.13 (5.10.0+dfsg-3build2, Feb 06 2018, 18:37:50)

I see no "pypy3" in bionic. I see a snap for pypy3, but only the edge version is up-to-date with code that can run, e.g., pandas (version >= 5.9)

I would prefer an apt package, and would assume that there would be more focus on pypy3 than pypy for python 2, but I'm either missing something, or that isn't true. What's up?

ubuntu-make (umake) can be used to install the latest version of a variety of popular developer tools.

How secure is this process and how does it compare to the security built in to apt-get, such as digital signatures via keys that are not stored on the repository servers, secure, automated updates when security vulnerabilities are identified in tools, etc.? Apt has an apt-secure man page with details on the apt approach to security. Does umake have anything like that?

Does umake, or the processes that create the packages it delivers, check any digital signatures on the underlying packages, including e.g. Maven Central signatures? How are the signature keys vetted? Does the process create any signatures that umake, in turn, automatically checks? These steps seem important, as discussed at Is Maven a plausible vector of attack? - Information Security Stack Exchange

I see that umake doesn't do updates yet (updating tools · Issue #74). Is there any way to determine whether a given installed tool is out-of-date, so you know when to do the "remove/reinstall" workaround? Is there any way to check installed umake tools for security vulnerabilities? Is the archive of packaged tools and any associated version and security metadata available for inspection directly over the web? If not, is it available via umake? What format are the packages and metadata in?

Finally, are there any plans to use The Update Framework (TUF) to really deal with software updates in a secure way?

Kotlin is gaining widespread attention as being a great programming language, and is now officially supported e.g. for Android development. But the options I see for installing it aren't clear about security issues. I don't know if sdkman checks (or even supports) signatures on packages, I don't know if it automatically tracks security updates on Kotlin and other installed packages (like apt-get does), I don't know how big the install is going to be, etc.

So to sum up

Is there an installation approach for Kotlin that is relatively secure (with package signatures, auto-update etc)?

E.g., is there a PPA for it? (Is anyone even working to package it for Debian/Ubuntu?).

Or does sdkman have the necessary properties?

Or is there some other approach?

Update: I see that ubuntu-make (umake) is an option. For IDEA and Kotlin, it seems the version from their ppa is still needed, as explained at Ubuntu Make 16.03 Released With Eclipse JEE And IntelliJ IDEA EAP Support, More

But I'm surprised that I can't easily find any information on the security aspects of umake, and it seems that they don't do updates yet (updating tools · Issue #74), so my question remains open.

The Go programming language (golang) has multiple versions and multiple implementations. Some software requires a more recent version than the 1.2.1 in Ubuntu 14.04 via the golang-go package:

$ go version
go version go1.2.1 linux/amd64

Or the version available with the gccgo-go package:

go version xgcc (Ubuntu 4.9.3-0ubuntu4) 4.9.3 linux/amd64

I guess that means it's based on gcc 4.9, which is a bit confusing but says:

The GCC 4.9 releases include a complete Go 1.2 implementation. The GCC 5 releases include a complete implementation of the Go 1.4 user libraries. The Go 1.4 runtime is not fully merged, but that should not be visible to Go programs.

Trying to install software using e.g. the go get command can be frustrating and unhelpful, generating core dumps or strange compilation errors.

What are the options for installing a more recent version? I need 1.5 or later right now, but would prefer a more general answer.

I don't want to just install a tar.gz package as suggested by the official documentation, since I want to get security updates etc. I'm actually alarmed to see that they recommend that.

I'm always nervous about installing from a ppa since they don't have security update guarantees, but that might be the best option, and https://launchpad.net/~ubuntu-lxc/+archive/ubuntu/lxd-stable seems like one option.

It seems that (at least in Ubuntu 12.04 Precise), the cgroup-lite package is used to provide some cgroup capabilities for packages like libvirt-bin and lxc. It sets up a number of cgroups in /sys/fs/cgroup for example.

I was reading the very nice documentation on cgroups for redhat and I wanted to use some of the other cgroup commands that are only in cgroup-bin. But when I installed that, /sys/fs/cgroup became empty, presumably because the cgroup-lite package was uninstalled since it conflicts with cgroup-bin.

What am I missing after having done this? Is there a way to get the same default set of cgroups that cgroup-lite sets up? Is there any documentation for the cgroup definition that cgroup-lite provides?

The lxc documentation in the server guide talks a bit about cgroup-lite, but seems to assume that cgroup-lite will be installed, which won't be true if either recommended packages are skipped, or the alt]ernative, cgroup-bin, is installed.

I've got a bunch of hour-long videos I want to split out into 5-10 minute pieces. So far I've simply used the openshot video editor to identify good splice points, and the blazingly fast avconv command-line program to copy out the actual audio and video without changing any formats or codecs. Then it is easy to upload that to youtube via the google command-line tool.

But I'm looking for something to help automatically find the right splice points. The video is mostly static slides with some real-time handwriting going on. So e.g. a script that would process the video and give me timestamps for the major video transitions (slide transitions) would help a lot, especially if it could then back up and go forward from those transitions to eliminate silence.

Is something like that available? Are other fast/easy video editing pipelines for this sort of work out there?

Running id3v2 -l file.m4a didn't find any tags, but displaying it in rhythmbox did. Is there another command-line program for working with tags that works on .m4a files?

I'm running precise 12.04.

Note: See Wikipedia's page for .m4a, .mp4, MPEG-4 Part 14 file format background, and note that as discussed at Spec for MP4 and M4A tags - Hydrogenaudio Forums, Apple considers the tags used by iTunes to be proprietary; hence there is no official spec for them available.

This, like other Apple/iTunes polices has caused no end of frustration for users working with audio encoded as AAC!

What is http://launchpadlibrarian.net (besides "a file repository used by Launchpad")?

What is there? How do you find out where things are? Are the urls permanent?

In particular, does it contain ppa .deb archives? Or is there some other way to get .debs for ppa packages?

I was surprised to see that since Natty 11.04, even when we have enabled a backports repository, updated packages aren't automatically installed from that repository. We have to install individual updates manually, e.g. via

apt-get install ipython/precise-backports

as explained at UbuntuBackports - Community Ubuntu Documentation. I guess I just didn't get the memo....

First question: how do we find out what our options are for upgrades via backports? I'm surprised that even if I ask for status, e.g. via

wajig status ipython

it doesn't tell me there is a new version available. I'd like a list of all upgrades for packages which I've already installed.

Next, how do I avoid losing existing packages just because I upgrade via a backport?

E.g. the above ipython install tells me:

The following packages will be REMOVED:
   ipython-notebook ipython-qtconsole

How do I say I want the latest backported ipython, as well as any other packages for which updates are available that depend on it, without manually figuring them all out and installing them also?

I know I can change the pinning so that I get all updates, but I'm hesitant to go against the general advice not to. But if a user asks for one package to be updated, wouldn't it be natural to update all the dependencies, like you'd get from a ppa?

Some event web pages have links to let you add the event to your calendar. They use a variety of techniques to do so.

When I click on, for example, Great American Sing-along, which seems related to the ActiveDataExchange product, using Chrome, it just offers to download a file "adecalendar.ics". Firefox offers the same option, and also offers to let me open the file with an app of my choosing, but only shows emacs as an option. And kitchen-sink though it may be, by default emacs doesn't seem to do any magic with .ics files ;)

I know that if I save it to a file, and go to https://www.google.com/calendar I can import a .ics file, but that involves a lot of frustrating steps.

So how can I just get a one- or two-click method of adding this to my google calendar?

Note that if the web site offers the calendar via a "webcal:" protocol, it is possible to configure a protocol handler so that gets passed off to the google calendar site, as documented at Google Calendar as default handler of webcal protocol for iCal/ICS files.

It used to be that when a program crashed, especially when a user was using a pre-release of Ubuntu, apport could be used to open a bug report. The user could then track the bug, see if it affected others, help fix it, etc.

As of Precise 12.04, this behavior and workflow changed. As I discovered in Bug #993450 “Apport fails to submit bug report”, by default apport no longer opens a bug report (and it is awkward but not impossible to get it to do so). At the same time people are noticing a new "whoopsie" process, as described at What is the 'whoopsie' process and what does it do?.

After some more googling, I dug this blueprint up, which describes the whole process: ErrorTracker - Ubuntu Wiki. (It didn't mention whoopsie or daisy, so I added them - please correct me if I got it wrong).

Wow - this sounds like great work to streamline and improve the crash reporting process.

I'm left with this question: how does a user learn what the status of the problem is? The blueprint now has this requirement

The user should have some way to check back on the status of their crash report; e.g. have some report ID they can look at to see statistics and/or any associated bug #. E.g. provide a serial number at time of filing that they can load via a web page later on.

which seems unimplemented. Is there anything available in the meantime?

And how does a developer get into the game? Going to https://daisy.ubuntu.com just provides an "Incorrect Content-Type" error message.

Finally, I suggest documenting the apport behavior changes in the Release Notes. It should be of interest to anyone who has been trying to help out Ubuntu.

The graph and bar chart at the Ubuntu Error reports page seem to contain a lot of information.

But I'm puzzled about the meaning of some of the values, and the page doesn't refer to any documentation.

• What does the "frequency" column measure, in what units?
• What does the "Mean Time Between Failure" vertical axis on the graph mean?
• What do the "If all updates were installed" vs "Actual" toggle mean? For me, clicking "Actual" just blanks the whole graph.
• Where is the code that generates the page?

Update: And where does the data come from? Is this related to ErrorTracker: how can I track a bug that caused a crash and was reported via apport / whoopsie??

I process lots of files via the command line, e.g. using my own shell or python scripts. Is there an easy way to add the files that I manipulate to the Unity "recent files" list so I can then easily access them via the Unity dash?

I'm trying to play the song referenced at A Prairie Home Companion: May 6, 2006 - Moonlight on the Colorado River lyrics. For some of my friends, the "listen" link there plays just the song whose lyrics are listed. But when I click on it (via Chrome on a Maverick system with Totem as my player), I hear the entire 2-hour program. The listen link redirects to a SMIL file which references "rtsp://mprrm.publicradio.org/5559/phc/2006/05/06_phc.rm?start=00:00:04:10.0&end=00:00:07:08.0", which specifies a start time of 04:10 and an end time of 07:08. But it seems that totem is ignoring the start and end times.

Are there open source music players in the Ubuntu repos that properly handle this link?

I'm guessing based on the How can I play .smil videos? - Ask Ubuntu - Stack Exchange that Ambulant might handle it, but I guess that isn't yet in the repos? And RealPlayer presumably works, but as pointed out at (RealPlayerInstallationMethods - Community Ubuntu Documentation), it isn't open source.

Update:

For Totem, the relevant upstream bug seems to be Bug 424708 – Add support for starttime, which is present in launchpad as Bug #605370 “.smi files are not played correctly” as @student pointed out.

For some insights on the relevant playlist format and editing see also Video playlists with start and end times - Unix and Linux

I'm still looking for some way to play these videos as intended.

The http://packages.ubuntu.com/ site provides great information on packages, with changelogs and much more, but only covers currently supported releases.

The rmadison program (devscripts package) will also give me command-line access to all versions of a package from currently supported releases (nice!):

$ rmadison gnome-mount
gnome-mount | 0.8~svn20080225-0ubuntu4 |         hardy | source, amd64, i386
gnome-mount | 0.8-2ubuntu1 | karmic/universe | source, amd64, i386

But I don't see a way to find e.g. what version was in intrepid, which is no longer supported. I know the intrepid release is available at http://old-releases.ubuntu.com/ but I don't want to have to go through each repository by hand.

I have a big USB disk with 3 partitions: one for backup and two other bootable ones for installing and testing new distros. I want the backup partition automounted on boot. But I don't want the two test partitions automounted. Despite my use of "noauto" in /etc/fstab, something (gnome?) seems to be mounting them when I plug the drive it.

LABEL=mybook            /srv/backup     ext4    defaults        0       2
LABEL=mybook-root       /media/mybook-root ext4    user,noauto  0       2
LABEL=mybook-spare      /media/mybook-spare ext4   user,noauto  0       2

In previous Ubuntu distributions it seems that it was possible to configure gnome so it would avoid mounting particular partitions on removable drives like USB:

gnome-mount --write-settings --mount-options noauto --device /dev/sda1

This is no longer available in Lucid (when did it go away?)

Is there another way to do this now?