Bug 399571

Summary:	[KNS] Update button can take a long time to get enabled when any backends are slow to update
Product:	[Applications] Discover	Reporter:	Antonio Rojas <arojas>
Component:	KNewStuff Backend	Assignee:	Aleix Pol <aleixpol>
Status:	RESOLVED FIXED
Severity:	normal	CC:	aspotashev, flareload, leinir, nate, oss+kde+bugzilla, s.scalabrino9
Priority:	NOR
Version First Reported In:	5.14.0
Target Milestone:	---
Platform:	Other
OS:	Linux
Latest Commit:		Version Fixed/Implemented In:	5.54
Sentry Crash Report:

Description Antonio Rojas 2018-10-09 18:42:28 UTC

SUMMARY
This is similar to bug 398746, but for the updates view. The "Update" button doesn't get enabled until all backends have finished checking for updates. If some kns provider has a bogus URL (like kalzium does), this can take up to one minute until it times out.

SOFTWARE VERSIONS
(available in About System)
KDE Plasma Version: 5.14
KDE Frameworks Version: 5.50
Qt Version: 5.12

ADDITIONAL INFORMATION

Comment 1 Nate Graham 2018-10-10 17:17:31 UTC

Not sure it would make sense to do an update before all backends have finished updating. Instead maybe we should focus on the underlying problem and improve the situation when a KNS provider has a bad URL.

Comment 2 Aleix Pol 2018-10-11 23:00:22 UTC

https://phabricator.kde.org/D16146

Comment 3 Aleix Pol 2018-10-19 19:05:40 UTC

*** Bug 399869 has been marked as a duplicate of this bug. ***

Comment 4 Simone 2018-10-23 13:13:56 UTC

Same problem here.

Trivial workaround: Removing the kns file from the system (e.g., /etc/xdg/kalzium.knsrc on Arch) or moving it somewhere else.

What about making KNS updates optional? I mean, all the KNS providers could be listed in "settings" with a checkbox to enable/disable them, just like other sources (e.g. Flatpak). If a provider becomes unavailable, they could be easily disabled from the UI (without removing system files).

Comment 5 Aleix Pol 2018-12-19 17:00:07 UTC

*** Bug 402287 has been marked as a duplicate of this bug. ***

Comment 6 Aleix Pol 2019-01-09 18:06:07 UTC

This issue was fixed with:
https://phabricator.kde.org/D17593
https://phabricator.kde.org/D17594

If it resurfaces we can decide to start timeout-ing backends.