https://bugs.kde.org/show_bug.cgi?id=516907
--- Comment #13 from flan_suse <[email protected]> --- (In reply to Kai Uwe Broulik from comment #12) > Use smb:// the which is proper asynchronous. > > As for finding the cause: a well timed Ctrl+C (SIGINT) while it's frozen > could identify the cause. I've gotten much better performance. better handling of characters, and server side copy with cifs/fstab. To use smb:// would be a pointless downgrade of performance for a new regressive bug from a recent update. I'm not going to change my workflow because of a new bug. I just did another test. I enabled all my SMB mounts, but had the shortcuts removed from "Places". For some reason, they all get mounted the moment I launch an instance of Dolphin, even if I don't visit the paths. (Yes, with a long delay before the Dolphin window is visible.) This doesn't just tell me "they're getting mounted", which isn't a problem. This tells me that all 14 paths are being entered. KIO is "doing something" for all 14 paths. Them being mounted is not the cause, but a clue as to what is going on. It's as if something with KIO polls available paths and wants to inspect them? Was there something new introduced into KIO Frameworks that polls or inspects paths whenever a KIO-associated application is launched/focused? The only way they are mounted is if I intentionally traverse *into* the path. But because all 14 are being mounted the moment I open up Dolphin, this means all 14 paths are being traversed under the hood without my consent. Using PCManFM-qt, the problem doesn't exist, and everything behaves smoothly, probably because it bypasses KIO Frameworks? -- You are receiving this mail because: You are watching all bug changes.
