Why are those notification even persistent in the first place? If a user misses the notification, they should notice that the job was finished by the fact that it's not running anymore.
Only errors should be persistent.

In D1478#27445, @colomar wrote:

Why are those notification even persistent in the first place? If a user misses the notification, they should notice that the job was finished by the fact that it's not running anymore.
Only errors should be persistent.

So that users may leave long-taking jobs unattended and then click "Open folder" on the notification or something once they're back and/or the job is done.

In D1478#27450, @mck182 wrote:

So that users may leave long-taking jobs unattended and then click "Open folder" on the notification or something once they're back and/or the job is done.

Ah, right, that makes sense.
Not grouping them can lead to a quite long list of notifications if someone does a lot of file transfer jobs, but that's certainly the lesser of two evils.
+1 for the change, then!

Please ignore the OSD stuff in the patch :)

In D1478#27445, @colomar wrote:

Why are those notification even persistent in the first place? If a user misses the notification, they should notice that the job was finished by the fact that it's not running anymore.
Only errors should be persistent.

i find the fact those notifications are persistent very useful, as if i have for instance a big file to upload i just abandon the machine.
when i return i prefer having something telling how the transfer went, i don't find the "no news good news" enough here

+1 from me(provided the commit excludes the osd stuff), would like Martin taking a final look at this

Ping.

Also, should we target Plasma/5.6 with that?