2020-11: why can't LibreOffice update automatically, still?

This one question has been asked many times over the years:

Why can’t LibreOffice (the software) update itself automatically, instead of nagging users with popups (which might include security fixes, or might not, the interface doesn’t say quickly) that say “update from 6.4.6.2 to 6.4.7 is available”, but “automatic updates are not available at this time”?

It’s November 2020 out there, the first questions about this seem to date as far back as 2012, and all users still have to download 300 MB manually, wait until it completes (300 MB on fiber is not short enough to not lose time, and not long enough to do go back to work on something else), click, do the windows authorization shenanigans, go through the screens, uninstall everything, and reinstall everything

The one thing Microsoft does better than LibreOffice is this. This is the single biggest pain point for me as a user, those nagging popups, way above anything else like 3D graphs or other. Kudos to the community for developing such complex software, but please, please, include this in the next release that is still open for discussion! I’ll donate if you do so! It’s not a whim and it’s not secondary, it’s been 8 years and it’s interrupting deep work and it’s breaking user experience!

People reading this who will also donate if automatic update is fixed, please reply or upvote. We have to do something, or it still won’t be fixed by 2030!

Another thought: is there some kind of “bounty pot mechanism”, where people commit money to an open source project, that gets into escrow once the dev team agrees to set a deadline on a specific feature, and gets released to the foundation only after they implemented the feature on time? I seldom participate in crowdfunding, but I would definitely give something on such a bounty. I know this auto-update feature would require a lot of work from the dev team, but such a mechanism might be able to provide the incentive to get enough developer attention to finally do it!

Edit: Following ajlittoz’s answer, if implementing the possibility to auto update is truly that complicated in this case, then a possibility would be to:

  • Sort the release notes by criteria set by users, and
  • Let users manage their criteria so they only see notifications for updates they decided could be important to them.

For example, my criteria would be:

  1. Is this update a security update, patching a vulnerability?
  2. Does this update prevent a crash on my platform that the developers think has more than 1 chance of happening in 50 000 hours of use?

If the update does not meet any of those criteria then I don’t want to see the popup; if it does, then by all means, yes. Currently, I have no way of doing this: either I turn notifications off (and risk missing a vulnerability), or I keep them on (and have to go read release notes for every update and upgrade publication)

why can’t LibreOffice update automatically, still?

Quite simple - nobody has/had the time, ability and/or will to implement this feature (and from my point of view: least important feature). You don’t seem to accept that the software is provided “as is” from a community of volunteering developers and donating is not paying a vendor. So the only solutions is: Get in contact and help to implement.


> The one thing Microsoft does better than LibreOffice is this.

… and Microsoft supports an ecosystem of hundreds of Linux distributions, right ?

Thank you for your comment Opaque. I did say I was grateful to the community for the software’s existence; I just wanted to find a way where the community could “democratically” voice its concerns and rank possible features by how important they saw them, to add better metrics to development decision-making. I can’t actually contribute code from where I stand today; I can only contribute to “bounty”-type schemes to voice what I would most like to see in future releases.

Also, I do not intend to say Microsoft is “good” (which it isn’t), just to underline that this popup problem was (a) important to me and (b) not even acknowledged by the developer community of LO. I will edit my question to incorporate the comment on Ajlittoz’s answer

First I’m no developer. Consequently my opinion is not an official one.

There are two issues in your question:

  • getting notified of an update/upgrade
  • installing the update/upgrade

Getting a clean notification of an update/upgrade availability is a great thing. It is good to know that perhaps problems can be fixed with the new release.

Automatic installation is another thing. Personally, I don’t want it because it may break critical utilities/applications on my system. I value reliability to the highest point, even if I experience problems with the current version of applications.

I prefer to check point my work, make the necessary backups and decide when it is time to update because I have spared enough free time to check things will be fine. Moreover, when I’m faced with an upgrade on a application which is critical for me, like LO, I prefer to test the new release on a sandboxed test machine before switching.

In my opinion, an update feature, whether automatic or manual, must not be part of the application but the target of some “side” utility. App developers have sufficiently hard time writing and fixing the app without the risk of shaking it through the addition of a feature which is rather on the system maintenance side.

Considering the multiplicity of systems on which LO can run, it would simply be insane to integrate such a feature inside the app itself.

I understand your frustration which is probably amplified by the fact you are under Windows. I made the “great dive” more than a decade ago for Linux. Under Linux, you have a packaging organisation where huge application like LO can be sliced into more or less independent parts or packages. This allows to install only the parts you are interested in. This reduces the download size well under 300 MB if you correctly target or limit your need.

Under Linux, package maintenance is under control of a package manager which may, on user option, run in the background to periodically check for updates. Another option, also user-selectable, allows or not automatic update. For reasons stated above, I never enable this one.

One last word. Package organisation and structuring is not done by LO developers. Let them work on bug fixes without overloading them with “packaging” and “delivery” issue. Linux distribution maintainers take the LO apps from TDF and then slice the apps according to their policy (this is why various Linux distros present different “slices”).

I find this two-tier organisation preferable though it causes some lag between TDF announcement and distro availability. Developers develop LO. Distro vendors care for user updates.

Once again this is a personal opinion not endorsed by TDF.

Thank you Ajlittoz for your detailed answer. With this in mind, I would suggest the best fix would be to find a way to (a) sort the release notes by criteria set by users, and (b) let users manage their criteria so they only see notifications for updates they decided could be important to them. For example, my criteria would be: (1) is this update a security update, patching a vulnerability? (2) does this update prevent a crash on my platform that has more than 1 chance of happening in 50 000 hours of use? If the update does not meet any of those criteria then I don’t want to see the popup; if it does, then by all means, yes.

Would that be practical to do?

In the original question, you objected to downloading manually. In Tools > Options > LibreOffice > Online Update under Download Destination you could always check the box that says Download updates automatically. You would still have to install it manually, but at a time of your choosing. Cheers, Al

BTW I hate the way MS installs stuff and wants to reboot when I’m in the middle of something important or urgent.

Regarding your question (2), there is no way to answer it. Even if your computer has its factory installation “unaltered” by any add-ons, as soon as your start customizing the system settings to fit the UI to your tastes, you introduce subtle modifications which make your PC unique. Consequently, it is impossible to state 100% in the release note that changes are totally innocuous. Of course, some changes are so light that they can be ruled out, but you can never for sure with the others.

@ajlittoz it seems you are referring to the risk that an update/upgrade might break the user’s system. Indeed, this is a very hard call. My question (2) was actually about the opposite: does the old version crash with many known bug reports corresponding to my overall platform (i.e. Windows 10)? If so, does the update specifically target this crash possibility and corrects it? If it does, then it has a chance of fixing something that might have crashed my system had I chosen to stick with the old version. Therefore, except for the risk you mention of doing more harm than good by updating/upgrading, I still should be decreasing the crash probability on my system, on average, by updating/upgrading. It’s this information I would love to see, e.g. in the first lines of the release notes, “fixed a system crash happening when doing XYZ”, and transforming the presence and relevance of such fixes into a switch that would show or hide the popup depending of the “relevance” setting I chose.

There are usually technical documents listing the bugs fixed in a release. This list is not included in the release note because a sequence of numbers does not mean anything to a common user. The release note mentions only significant fixes, those which are user-visible or impact the usage (=method of use) of the application.

Usually, you’re right, a new version improves reliability.

However, I don’t want to risk a hiccup and prefer to remain in control when working on critical documents. I decide when to update.

Your biggest inconvenience, as a Windows user, is the monolithic aspect of upgrade: the full application and its options are downloaded instead of the pieces related to the fix. Full replacement creates the risk of not distinguishing between update and initial install, i.e. erasing your customization.