Created attachment 135613 [details] Duplicate installed application Fresh Kubuntu installation. Discover shows two "KDE Connect" apps installed, and other duplicate apps can be found, such as Deluge.
Created attachment 135614 [details] Duplicate apps among available ones
Created attachment 135627 [details] qBittorrent There are two qBittorrent apps, with the same name and almost identical logos. Maybe one a Flatpak and the other is a snap, I can't tell. But if that were the case, why don't they have specific names, and why are the logos slightly different? Or maybe one is malware, trying to benefit from the original brand.
"KDE Connect" and "KDE Connect Indicator" are different apps, so it makes sense that they have separate entries. The fact that there are three Deluge entries is because: 1. we match package names too, and the exact package name is "deluge", so you see an extra entry. This is already fixed in Plasma 5.22 in Bug B 2. The app is available from multiple backends (e.g. distro package and also Snap/Flatpak), and unfortunately Discover can't de-duplicate them because each entry provides a different AppStream ID. The display of this is improved in later versions, where we show the icon of each backend so you can tell the difference. See Bug 399530. This is the same issue with qBitTorrent. So thankfully, all of these issues are already fixed in newer versions! Phew.
Created attachment 135682 [details] Fixed screenshot for KDE Connect duplicate
Nate, thanks for clarifying (and for looking at all my bugs!) My bad on KDE Connect/Indicator. I've just fixed the screenshot. I'm using Kubuntu 20.04 LTS for stability, but perhaps I should switch to Plasma 5.21 on the 16th? What kind of stability should I expect? (I don't mind functionality bugs as long as the system overall is stable, and 5.18.5 is).
Thanks. I strongly suspect that it's the same issue with KDE Connect. In general, I encourage KDE users to avoid LTS/"Stable" distro releases. KDE's software is updated on a very rapid schedule, and in general you get a better experience by sticking with the latest versions of things. You get more bugfixes and UX improvements than you get regressions, in my experience. You can upgrade to Kubuntu 20.10 which has Plasma 5.19. Kubuntu 21.04 will have Plasma 5.21 in a few months when it's released. If you want to track KDE schedules more closely than that, your options are a rolling release distro like Arch or openSUSE Tumbleweed (my current distro of choice), KDE Neon, or a discrete release distro that upgrades KDE components regularly, such as Fedora.
Created attachment 135720 [details] KDE Connect confusion Thanks for the tips! Back to KDE connect, the confusion stems from one of the two launcher entries in this last screenshot being the KDE Connect *settings* module, and the other being the KDE Connect application. If you can't easily tell which is which, then that's the bug :)
You're welcome. I agree that's confusing. However the data for those entries isn't in Discover; it just consumes that information and displays the result in the UI. The solution would be to improve the AppStream Metadata in KDE Connect. This would be a great first patch. Would you like to work on that? I can help you!
Thanks for the offer, Nate. It unfortunately looks like upgrading from 5.18.5 to 5.21 brought more regressions than bug fixes. I've tallied up the main ones at https://www.reddit.com/r/kde/comments/ll388r/plasma_521_is_out_and_this_time_we_have_improved/go7koib/, and I'm out of bandwidth at the moment to delve into figuring out the whole development setup and patch submission process. If the code were hosted on GitHub, with in-browser editing, that *would* be below my PITA threshold*. Besides that, I'm not sure that updating AppStream data would be enough. What about the different ratings and reviews that have accumulated for the different versions of the same app? Take Spotify, for example - there's a difference of about 100 ratings between the two versions that Discover shows. Honestly if bug 433166 keeps happening, I'll have to revert to 5.18.5, which I know was stable. * I've articulated way back in 2009 why GitHub's superior ease of use would favor any projects hosted there at https://web.archive.org/web/20200626124617/http://wiki.dandascalescu.com/essays/pita-threshold. PS: I'm not sure the status of this bug is really "FIXED", given the discrepancy in ratings.
FWIW our code is hosted on a Git*Lab* instance, which has web editing functionality.