https://bugs.kde.org/show_bug.cgi?id=454946
--- Comment #33 from Aaron Rainbolt <arraybo...@gmail.com> --- OK so that didn't work. I tried to diff between what I thought was a working point on the master branch, and the tip of the Plasma/5.26 branch, and got a diff that was, I kid you not, nearly two million lines long. So. Looking at the backtrace Erich provided, I noticed this: #13 0x00007f2c8e4e808a in Plasma::Applet::destroy (this=0x556422a17b90) at ./src/plasma/applet.cpp:309 #14 0x0000556420a68254 in ShellCorona::unload (this=this@entry=0x5564227b7b20) at ./shell/shellcorona.cpp:834 #15 0x0000556420a6bdd8 in ShellCorona::unload (this=0x5564227b7b20) at ./shell/shellcorona.cpp:606 #16 ShellCorona::loadLookAndFeelDefaultLayout (this=0x5564227b7b20, packageName=...) at ./shell/shellcorona.cpp:602 The fact that ShellCorona::unload gets called twice in a row is very suspicious to me, so I'm going to try to reproduce the backtrace with Plasma/5.26 and see if it's still happening, then hope that the line numbers lead me to the right spot. If that fails, I'll try and find a "bad" spot on the master branch. -- You are receiving this mail because: You are watching all bug changes.
