https://www.youtube.com/watch?v=uX2mopyE5Z0


Bueno pues hemos desactivado el indexado de archivos y carpetas ocultas y activado el indexado del contenido de archivos. 6 GiB menos de caché de archivos.
No es suficiente teniendo en cuenta que si ejecutas el drop de la caché ésta se queda en 20 GiB menos (14 GiB si descontamos los 6 GiB mencionados). Linux mantiene esa memoria ocupada por las risas.
So #KDE filemanager #dolphin doesn't search through symbolic links and I had already spent about a year or so cursing #baloo whereas it worked fine on my laptop.
I had to get #fstab bind mounts to mount specific directories from another drive onto my home directory. Everything worked!
Now using #zfs meant I had to force the correct order of mounting else I have empty directories.
fstab's #systemd option - x-systemd.after=datapool.mount came to rescue and everything works again.
Creo que lo del aumento del tamaño de la caché de disco es por haber activado #Baloo. En cualquier caso pésimo trabajo por parte de Linux al liberar memoria para que la usen procesos que la necesiten. #Kswapd tarda mucho. Es más rápido y efectivo indicarle a Linux que escriba todo a disco y libere dicha memoria: sudo bash -c "echo 3 > /proc/sys/vm/drop_caches".
No quiero swap con un SSD. Gracias. El trasiego («thrashing» en inglés) gratuito no me entusiasma.
Con la cantidad de RAM disponible hoy día el uso de swap sólo retrasa lo inevitable: matar procesos o aumentar la RAM.
Question to the #KDE #Plasma developers out there: Did anyone ever think about building #OCR into the file indexer? I would love to be able to find screenshots based on the text they contain 😅
I found a thread on Discuss but no-one ever replied to it:
Does Baloo - KDE Community Wiki (and by extension Plasma/Krunner - KDE UserBase Wiki) support (anything similar to) Classifying images based on the text within them - Microsoft Community Hub (for .TIFF files at least)? I ask because I have a lot of .TIFFs with text in them (albeit not searchable, because unlike Searchable PDF there is no standard way to add it to my knowledge) so I’d like to be able to index them using OCR.
A #desktopenvironment has to work properly out of the box.
If it includes cpu and storage killers like #baloo, as the desktop environment #KDE does, then it is a k.o. criteria. Then you shouldn't use it.
Dear people from KDE: please remove baloo from KDE for the time being. It's not an essential feature, and it only causes problems atm.
You can fix it, but you shouldn't include it in this broken state in "normal" releases.
Ich glaube ich muss #KDE wieder löschen/deinstallieren von meinem Laptop. Das ist ja furchtbar, wie viel CPU-Leistung und Speicherplatz #baloo braucht?!?!
Mehrere Gigabyte Speicherplatz und 100% CPU-Auslastung für >15 Minuten, der Lüfter lief auf Hochtouren.
Und das nur zum Dateien indexieren und durchsuchbar machen.
Und baloo alleine entfernen geht irgendwie nicht, weil das zur KDE desktop environment offenbar dazugehört?!
Dann doch lieber #Gnome.
I've been using KDE Plasma for a few weeks now (because for some reason my framerate in Gnome Shell keeps getting worse with each new version. Yes, so bad it's noticeable)
And seriously: what the hell is baloo_file_extractor doing? It's just always there, wasting on average 20% of CPU. Extracting/indexing what? I have no idea.
The File Search settings tab claims the file index is "96% complete", but that it's also "Not Running", while `balooctl6 status` confirms it *is* running.
I also have no idea if this thing has sane defaults, or if I should add paths like ~/.cache, ~/.thunderbird, ~/.gradle, ... to the excluded folders. I have no idea, all I know is that I did do that and it made no difference