Summary: | Continue no longer the default when deleting items in split | ||
---|---|---|---|
Product: | [Applications] kmymoney | Reporter: | Ian Neal <iann_bugzilla> |
Component: | general | Assignee: | KMyMoney Devel Mailing List <kmymoney-devel> |
Status: | RESOLVED UPSTREAM | ||
Severity: | normal | Keywords: | regression |
Priority: | NOR | ||
Version: | 5.0.2 | ||
Target Milestone: | --- | ||
Platform: | Fedora RPMs | ||
OS: | Linux | ||
Latest Commit: | Version Fixed In: | ||
Sentry Crash Report: |
Description
Ian Neal
2019-01-12 14:04:33 UTC
It works for me with using KDE Frameworks 5.45.0. Could it be due to a change in upstream libraries since you use 5.53.0? The documentation at https://api.kde.org/frameworks/kwidgetsaddons/html/namespaceKMessageBox.html#a9a7aa0faa2cf861076586069472d11dc states that the Continue button is the default. The behavior is completely controlled by the KMessageBox object. KDE Frameworks 5.53.0 is the latest version that comes with Fedora 28 (and Fedora 29), not sure what other information I can provide. I can see from the KDE Frameworks code that if KMessageBox::Option is set to Dangerous then the cancel button has focus but not sure what is causing that in this case. It is happening in all ContinueCancel dialogs, not just the one for deleting a split item. Also tested using 5.0.2 RPMs from Fedora on Fedora 29 on a different computer with a new blank file. Every dialog that you would expect the continue / yes button to be the default is not, so has the same issue. Tested on another computer using the distro's KMyMoney running: Ubuntu 18.04.1 LTS Plasma 5.12.7 KDE Frameworks 5.44.0 Qt 5.9.5 xcb windowing system Issue happens on that platform too, so not just an issue on Fedora. Was still happening under Fedora 29, but under Fedora 30 now back is it should be. In Fedora 30 the SOFTWARE/OS VERSIONS are: KDE Plasma Version: 5.15.5-1 KDE Frameworks Version: 5.64.0 Qt Version: 5.12.5 I'll mark as resolved upstream unless someone has a better suggestion If the fix (and original cause) was in Frameworks, as it seems, then I agree, upstream. If the issue was in the Fedora packaging, it could be downstream. Hard to tell if nobody else noticed, and not worth further effort. |