There is no point in beta testing something if you donât know what features youâre beta testing.
Kinda like riding a bike. You need to know what parts are on the bike in order to know everything is working right.
There is no point in beta testing something if you donât know what features youâre beta testing.
Kinda like riding a bike. You need to know what parts are on the bike in order to know everything is working right.
Maybe the point could be to check the features and interfaces you are using already? Or apps that could possibly be broken because of changes in the core?
Itâs become a bit of a trend with nextcloud. The release version is riddled with bugs so people donât upgrade until the third or fourth patch.
Itâs riddled with bugs because the betas donât list what features need to be tested and they donât spend enough time in beta.
You want people to beta test but you donât tell them what theyâre beta testing or what needs to be checked. And thus, nobody wants to beta test for them.
Iâve never found withholding information to be helpful when you are trying to get cooperation and build community.
Just an idea: what if they/we tried it once? Just share the patch notes along with the beta release and see if everything goes better or worse, if people engage more with the content or not, if more testers sign up or less - that kind of thing
Just my 2¢
Many here have a philosophical disagreement with NCâs testing approach. In effect, NC is asking users to regression test beta releases to see if what they routinely do is still functional. But that doesnât test the full integration if users donât stumble upon the new functionality and exercise it to see if the normal use breaks after the new features have been put in use. It also crippleâs NCâs best chance at critical feedback by not informing sophisticated users what to look for as possible concerns when beta testing. Just because there is a new dog in the facility doesnât mean the dogs play well together if nobody opens the door to the new dogâs kennel. Reports from people that donât know there is a new dog or how to get to the new dogâs door are false indications of how the dogs get along.
In the past, they gave away a few details:
There have been a few discussions recently about testing, however not much feedback from Nextcloud. And from previous bugs, we donât know if there were analyzed and their internal testing was adopted to catch similar problems in the future.
In the first releases, my incentive for testing was to satisfy my curiosity about new features. It gives some time to check them out, find possible problems etc. before running in productive environments.
Itâs funny. If it wasnât sad at the same time.
What am I talking about? Well the community strongly demanded open betatests⌠So NC offered open betatests
Then the community loudly yelled for betatests that were announced so that everyone got to know about those. So NC announced betatests with a downloadlink.
And now ppl of the community are: well if we donât know whatâs new we wonât betatest.
Someone from above was referring to so called "sophisticated user"s⌠I daresay if a user is really as sophisticated they will find out about new features themselves (since they would be sophisticated enough to check github for informations).
So I think this whole discussion about new features or not is ridiculous. Go betatesting. Or just leave it. But please stop complaining about everything all the time.
You see the same effect on HackerNews, where certain headlines draw certain people to read and rebut certain articles, and it seems like all of HN is on one side of a debate one minute, and the opposite whenever the next article shows up. Thatâs an illusion.
The truth is this happens because the community includes a lot of different people. People with requests and complaints about this might not be the same individuals that criticized something else.
People have different needs, and thereâs some push-and-pull over whose needs should have the most resources dedicated to them, but the important thing is that this isnât always a zero-sum game. People âstrongly demandingâ open beta tests, âloudly yellingâ for announcements, and now asking for details to what to test arenât in conflict, theyâre all asking for a more transparent process; one that can meet the most would-be contributors (and yes, testing is contributing) where they are.
For folks who arenât familiar with GitHub (or donât want Microsoft tracking everything they do there!), commenting on the forum serves the same purpose as filing an issue. Every complaint you see here is from someone who feels the process could be streamlined, so that the Nextcloud project could benefit from the most contributions. Thatâs a good goal, even where a different method of communication might be preferable.
as someone else already said above: NC wants to keep itâs real news (ânewsâ in like âwhat is newâ) closed until the official release day.
and as I said above: ppl interested in that could go and check github (if they donât like to be tracked they could use tor or some vpn or whatever) beforehand for what they want to know.
apart from all of that: This âthreadâ is just an announcement about an betarelease. nothing more. Everything weâre commenting here has nothing to do with it.
The goal is construction, so please take this as the above in the spirit itâs given. Friendly debate and advocation, and not an attack on anyone.
This is rather the point. They donât want to keep this stuff totally secret, or they wouldnât even have code on GitHub prior to commercial release. They want to be able to manage the hype cycle, sure, but they need people to test these things.
So when betas are announced as ready for testing, but NC doesnât tell people what is new, folks like tflidd helpfully post the link to GitHub, and point people in the right direction. Thatâs part of the development process. The unofficial development process, but still very real. The rest of the thread has been a discussion on whether and how to improve this process, perhaps by making it official. Managing the hype cycle doesnât necessarily require silence on new features in the post where they are announced as ready for beta testing.
This discussion could and has occurred in other release announcement threads, youâre right. But since this is a process change and not a code change weâre talking about, is there a relevant GitHub repo where we should file an issue to request the change instead? Would starting a separate thread on the forum be more appropriate or impactful?
This is mostly an aside, but to provide a bit more context: The tracking issue isnât about whether we can see the code on GitHub, but whether we can participate in conversations (and file issues) there. [âŚ] Edit: Forget I mentioned GitHub or tracking. Participation has always been the point, which is why we even have these forums and calls for testing.
There is no âenoughâ you can do so that nobody will ever complain, but thatâs ok. A project with no complaints is a project that nobody uses.
the other second it just was getting to know about the new features in a new version.
So NOW that isnât enough anymore and everyone wants to take part in the discussion.
And again: itâs never enough. Like I pointed out above. And so this went round full circle: It is never enough.
Somebody always tries to have the last word on something that doesnât really need any more discussion.
If you wanna betatest you know what to do. If not itâs ok as well. Just donât try to find zero-arguments just for the sake of finding a contratry point of view.
The interesting part with the testing is, how do you get the right people start testing. Itâs not the most interesting thing to do. And having 10 more default debian/ubuntu environment that updates without problems, wonât help so much either. Or isnât it the update and environment itself and rather the UI, translations, âŚ
Imagine a list of 100 points to test, you have 10 people that test the first 10 points of that list. Iâd prefer to split them, that each focuses on 10 different points. However, if there are 3 people in the end that are testing such a beta, thatâs not worth the effortâŚ
All the bad bugs that were missed during the last updates, is that something that is already covered by some internal process, are there platforms less or not tested which often create problems? Is that something we know?
We all want to have a better initial releaseâŚ
Upgrade went smooth. Have some indices to update.
Got one error in the log so far. Will test more
PHP Required parameter $logger follows optional parameter $principalPrefix at /var/www/nextcloud/apps/dav/lib/CalDAV/CalendarRoot.php#37
2 posts were split to a new topic: Bug Hunters group created. Volunteer for testing and reporting breakage
Why is that a facepalm?
How can people beta test if they donât know what theyâre beta testing? Are our devs all from Microsoft?!?!
a perfect lesson on how to NOT pick any quote out of itâs context.
Maybe everyone who is willing to test should just test his personal business/use case, because that would help you to identify problems for your own environment, caused e.g by a regression, and therefore already provide a very good quality assurance, and as a bonus you would surely stumble upon one or the other changes such as new features and functionalities yourselfâŚ
Upgrading to beta 1 went with only one error. Now we already have beta 2 Tried that but got a message on extra files.
I understand that this is a common issue. I find lots of posts in the forum.
I think there is a place for a more detailed âHow toâ or a passage in the manual about how to solve this common problem. And at a level so a beginner on Linux can follow. As NC is used by a lot of amateurs like me I think more basic instructions would be appreciated and increase the usability.
After sudo -s I could cd to /var/www/ and identified the file and folder.
Should I just delete the file (rm?) and the folder (rm -r)? Or should I move them to another place?
For the moment I solved it by returning to my snapshot.
I now upgraded to the Beta2 but this failed with error
InvalidArgumentException: Column âoc_bookmarksâ.âurlâ is type String, but exceeding the 4.000 length limit.
And while complaining about unclear instructions. When I wanted to add the missing indices I tried the instruction given in NC itself to run âocc db:add-missing-indicesâ but although I was root and issued the command in the /var/www/nextcloud/ I got an error message Command âoccâ not found, but there are 21 similar ones. So I needed to run the command âsudo -u www-data php occ db:add-missing-indicesâ. I would propose that you give this instruction instead.
I am a fan of NC but I think a more client oriented information would help a lot. Many users are amateurs.
I must honestly say that I donât quite understand the idea behind this strategy either and I also doubt that this is the right strategy. But in the end, it is the decision of Nextcloud GmbH how they wanna handle thisâŚ
Personally, I donât care much, because I have simply adapted my strategy of how to deal with upgrades to the strategy of how Nextcloud GmbH is dealing with their announcements. This means that my personal âbetaâ testing takes place between the initial release and the first point release, which is perfectly fine to me, given Nextcloudâs fast release cycles.
Amateurs shouldnât deal with beta versions at all. And if they want to learn and participate, they should run a separate test instance. Matter of fact everybody should run a seperate test instance. Also there are many appliances like NCP, Nextcloud AIO, the Snap package etc out there. And there is a reason why most of the appliances are lagging a bit behind with upgrades of the dependencies and sometimes even with Nextcloud versions. If youâre an amateur and just want to use a product, you shouldnât use bleeding edge or even beta versions. Use either one of the appliances or if you want to install it manually use Debian stable or Ubuntu LTS and the PHP, MariaDB releases from the repositories of the distro. Also do not upgrade Nextcloud before the first point release.