I’m still amazed how (long term) plasma users can miss the cache cleaning and settings reset when problems arise after upgrading plasma (releases).
3 Likes
Some long term plasma users live in their own isolated worlds and only pop up now and then, seeking some wise council
4 Likes
In their defense, it shouldn’t be necessary. No software worth its salts should expect its users to clean the cache (and possibly some configs) on a regular basis.
It’s akin to expecting internet customers to power-cycle their modems every month or so to maintain their speeds and reliability. (This is not a joke. Without divulging too much, this was actually told to me by tech support of a large ISP, as if it’s “normal”.)
4 Likes
That’s a fair comment.
That doesn’t help anyone: reality shows that it is. Arch wiki explicitly mentions it.
Agreed and easy to implement.
Very difficult to do even with simple configs for a whole lot of reasons. Configs would have to be versioned and migrated, which in itself presents a whole lot of problems. Not everyone (in the sense of software-user which would be the distros out there using KDE) updates/rolls with the same speed - the config-migration code would grow extraordinary to accomodate all version jumps.
Very few software projects attempt this at all (for configs), I can name none of the top of my head.
Another practical problem arises: no volunteer builds such things and keeps amending it for each change (remember: KDE is mostly developed by volunteers).
I disagree: your example points to fear of memory leaks which power cycling would “keep in check”.
It doesn’t quite apply here as the software on the modem is always the same but somehow faulty.
For KDE users with said problems the software doesn’t stay the same: it changes (constantly on rolling releases like Manjaro).
I think, a more fitting analogy would be: You sand down the wooden cabinet before you give it a new coat of paint. If you don’t, the paint will eventually peel off due to the many (previous) layers.
1 Like
I didn’t mean it as an analogy to the software world, or KDE in particular. I gave that example from a user-experience perspective.
The user should only be expected to use the product. Going out of their way to research and discuss in community forums should be an exception to the rule.
Having to take matters into your own hands to create new config files or clear caches is not a positive user-experience, nor would I expect it from anyone. It’s an excuse for the developers not to address an issue. No software is expected to be “hand-held” through updates, even major upgrades, as it should gracefully ignore old configs and/or obsolete parameters.
How often do you hear people resolve their Google Chrome issues (on any platform) by resetting and reconfiguring their browser’s profile after so many upgrades? A web browser, like Chrome, is in essence “rolling” release software.
“Just backup your bookmarks, and make a note of your preferences, then delete your configs and let Chrome re-create your user profile, then just re-import your bookmarks and redo your preferences. It’s expected you should do this every five updates or so…”
The Chrome developers would be blasted by their users if they actually tried to defend such behavior.
I know we’re not “customers” of a product in the sense of paying for a good or service.
I know we can’t simply make demands of the developers “just because”.
I agree with that.
However, bad software design is bad software design: doesn’t matter if it’s free, paid, open-source, or proprietary
If your software expects your users to manually intervene by deleting and resetting config files as well as clearing the cache as new updates come out, then it’s a sign of poor software design.
There are many examples of applications and software suites that don’t commit seppuku because the user didn’t clear the cache or delete their config files. 
A particular grievance I have with the advice to outright delete the entire ~/.cache directory (as written in the Arch Wiki’s KDE page), is that it can seriously negatively impact a user.
Thumbnail previews are cached inside here. By wiping them entirely, browsing directories that contain many images and videos will have to re-generate the thumbnails, as well as create needless extra writes and fire up the CPU. This is especially a bummer for those who store tons of multimedia on external drives.
Then there’s applications that use the cache to immensely speed up their operations.
Within seconds I can check for duplicates across 20,000 files using Ckzawka, since it saves a cache of checksums from previously inspected files that share the same exact filesize.
Without this cache, the program would take very long and tax the CPU to re-generate all the checksums all over again.
User-experience is precious.
3 Likes
Same here! Random crash on 5.26.3 and sometime plasmas his unresponsive when getting out of sleep using a 32gb of ram 
But I’m fairly certain alot of users are given exactly 
advice when facing problems requiring manual intervention on Manjaro …
I’m not a chrome user and don’t read such topics. I’ve seen that advice given to firefox users sometimes though. There seem to be some bugs manifesting in borked profiles 
Everytime you see a pacnew you’re expected to manually intervene:
This is a line of thought / way of life on all arch related distros.
Agreed. To give them some credit, there is a warning:
Optionally, empty the
~/.cache/folder contents, however, this will also clear the cache of other applications:
As I understand the “wiki” concept though: you’re free to make suggestions to better the content especially if you’re seeing problematic, no longer valid or outright false statements.
Methinks, that is also why it’s always better to try to reproduce the problem with a fresh user (who would have no cache) than to clear (all) cached information.
Any advice is only that: advice. It doesn’t have to be taken literally especiall if you know what you’re doing. If you’re technically skilled and confident enough, you’re surly able to identify contents in ~/cache which be left out when cleaning up.
Another way is simply to (temporarily) rename ~/cache: you won’t loose anything but can test if cached data was involved in some problem you’re experiencing.
It’s your system after all.
1 Like
Just a random thought:
Instead of clearing the whole cache on every update, wouldn’t it be possible, or better to clear the cache per-application? For example:
- If Firefox is updated, remove the Firefox cache for all users;
- if Chrome is updated, remove the cache for Chrome;
- if Vivaldi is updated…
- etc. etc…
I don’t know how easy it is, but it might be, and I think it’s not too bad of an option.
1 Like
I don’t know about Vivaldi, but Chrome, Chromium and Firefox maintain their cache outside of ~/.cache, and so did Discord — I’ve recently removed that package, and it had about half a GiB worth of cache in its designated directory under ${HOME}.
Yup. I believe they’ll each have their own. However, that doesn’t really change my suggestion. 
1 Like
IMHO that’s actually not too bad.
$ du --max-depth=1 -h ~/.cache/mozilla/firefox | sort --human-numeric-sort
1.1G /home/mirdarthos/.cache/mozilla/firefox
1.1G /home/mirdarthos/.cache/mozilla/firefox/31vh7rnd.dev-edition-default
2 Likes
I’ve a series of bugs and issues coming from 5.24.x. Some could be related to kernel updates from what I understand. If any have already been mentioned then apologies. I had none of these issues on before the recent updates.
- On startup the progress bar stops at about 80%, then a minute or so later I get a black screen with cursor and then 20 seconds later the desktop is shown. On 5.24 there were no freezes, from login to desktop took about 2 seconds.
- Yakuake shortcut key doesn’t work at first. To get the shortcut key working it’s necessary to launch Yakuake, close, relaunch again. Yakuake is present in autostart.
- Pipewire sound randomly disappears. Seems to relate to Easy Effects as opening Easy Effects restores sound.
- Attaching keyboard sometimes simply does not work. USB delivers power as num lock is lit.
- Various stability issues and graphical glitches which seem to freeze the window manager but none that appear consistently.
Any suggestions welcome.
Well lets face it. Plasma has some migration scripts for older configs, but sometimes they fail as the user added their own stuff. Customization is not wanted these days, even on KDE space. As soon as you derail from the default settings and add your own tweaks those migration scripts may fail as they are only tested on stock settings. So you can upgrade from 5.24 to 5.25 without issues and then to 5.26. A jump from 5.24 to 5.26 should also been possible.
However, only if you used the standard settings and breeze theming. Our Breath theming is more or less a slight alteration of the stock Breeze by now. Just re-colorization so to say. KDE software went so complex that upstream can’t test all those settings a user can do. At some point some features got dropped as not many users used them. Those who did the experience broke.
From what I read: 5.24 LTS was a good release and any feature release who followed broke here and there.
The future of Plasma 5
Plasma Frameworks reached end of development with 5.101 series It is the last KF5 release with feature development, there will be more releases after that for maintenance (e.g. bug fixes), but with increasing intervals as the amount of changes decreases. This might add some extra stability and the opportunity to actually fix given issues in frameworks, which is the base for their Apps and Plasma.
Nate is blogging what is going in KDE world. So you can know what kind of thing will hit the fan next before you update any KDE software. Fixing bugs: well the 15-Minute Bug Initiative started out with 100 bugs, and 11 months later we’re down to 47! There have been 142 15-minute bugs, and they fixed 95 of them, for a fix rate of 67% in a year. Wayland fractional scaling and multi-screen got fixed. That is something to look forward to, but I doubt it will be backported to older release series. Their 20k fundraiser is still ongoing. So far they collected about 13,7k €.
So we will see how stable the last 5.27 LTS will be, before Plasma 6 cycle is going to start. There is no Schedule yet but you can look at some project boards. Things missing on Qt6 can be found here.
When the focus shifts to Plasma 6 development getting fixes still into the old 5 series will be harder as KDE developers are not so many.
away out …
So if there is some lookout for stability, maybe XFCE 4.18 might be a good thing. You can give it a full swing with today’s testing build. Especially the changes to Thunar filemanager are nice. And there is the Gnome desktop which is more organized in a more corporate structure by Redhat and Co.
4 Likes
XFCE! As I use xfce Vanessa 4.16 in my “Netflix” laptop, I will be assessing it from every possible angle. It has been SUPER stable.
(As my Netflix xfce lappy uses Thunar natively, I decided to check it out for my KDE system. Split screen finally. Good job!)
1 Like
Maybe this is why I’m never running into trouble when upgrading KDE, I always run default Breeze and neither of the mentioned updates has caused me any problems. Good thing I like Breeze. 
BTW, that XFCE changelog, especially for Thunar, looks very tempting.
4 Likes
Good on you. I’m running Plasma 99% stock standard. And I’ve had some niggles and one or two near heart-stopping moments when you feel…very…very…alone all of a sudden
Keep crossing those fingers 
PS, there is life after a Plasma Crash!
I just installed the new Thunar on my KDE system. Very nice! Easy to use!
1 Like
How big were the dependencies downloads?
You know it would take you one second to know by yourself?
[omano@omano-nvme ~]$ sudo pacman -Syu thunar
:: Synchronizing package databases...
core is up to date
extra is up to date
community is up to date
multilib is up to date
:: Starting full system upgrade...
resolving dependencies...
looking for conflicting packages...
Packages (6) exo-4.18.0-1 libxfce4ui-4.18.0-1 libxfce4util-4.18.0-1 startup-notification-0.12-7 xfconf-4.18.0-1 thunar-4.18.0-1
Total Download Size: 2.56 MiB
Total Installed Size: 15.06 MiB
:: Proceed with installation? [Y/n]
1 Like