If we will use media-video/kffmpegthumbnailer or kde-base/mplayerthumbs and point mouse at some video (with dolphin) that is on removable media device state will be set to busy, and umount will be locked. Now if we would like to umount device we will get message that device is busy, even if we will point other file that isnt on device and new thumbnail will be used in dolphin, thumbnailers will still use and block that device.
Steps to Reproduce:
1. Mount some removable media with video file
2. point video file with mouse
3. try to umout device
Device not unmounting if we will point video file on dolphin
Device umounting when we want, not when kde will let us to do so.
kde-base/mplayerthumbs should be removed from meta dep of kde to time when this bug will be fixed.
Which kde version do you have?
Its problem of all kde versions that are in portage (now i have 4.4.2).(In reply to comment #1)
> Which kde version do you have?
Its problem of all kde versions that are in portage. Answer is 4.4.2 (right now) if you realy need to know.
This bug dosn't affect ssd cards from my
08:06.1 SD Host controller: Ricoh Co Ltd R5C822 SD/SDIO/MMC/MS/MSPro Host Adapter (rev 19)
I tryed them and they are unmounted on the fly without any problems.
I can confirm this bug, it's very annoying.
Question in my mind is, how do we know where the bug is? Is it in the actual media thumbnailer, or in KDE's kio_thumbnail process? My guess is the latter, since when I list running processes, there is no "ffmpegthumbnailer" (in my case), only kio_thumbnail. My additional guess is that kio_thumbnail is keeping the device open so that it can quickly provide thumbnail images as the user requires, which is fine for a fixed device, but problematic for removable devices.
Could a KDE expert comment? My last and final guess (really) is that we'll have to take this upstream. :)
Btw, here is the upstream bug, which was lodged more than two years ago. :\
Waiting for upstream... ... ... ...
According to upstream report, affects also 4.4.5
But the important question probably is, does it affect 4.6.0?
(In reply to comment #8)
> But the important question probably is, does it affect 4.6.0?
No upstream duplicates since kde-4.4. So I'll tentatively assume that this is fixed...