https://bugs.kde.org/show_bug.cgi?id=387006
--- Comment #11 from Christophe Giboudeaux <christo...@krop.fr> --- Git commit a38e639ad913d1dce1f808029333fc00995b641e by Christophe Giboudeaux, on behalf of Harald Sitter. Committed on 13/09/2020 at 10:32. Pushed by cgiboudeaux into branch 'master'. simplify qrc watching test This test was a less sufficiently tooled variant of the existing removeAndReAdd in terms of watching things apper and disappear, which the test asserts does not break when one tries to watch qrc paths. While I get the idea of wanting to have a very truthful test case in our case we have to take reentrancy over reproducibility since the test may be run multipel times in paralla. I've verified that undoing the qrc escape in addEntry() actually still breaks within the context of removeAndReAdd and so this is a 100% equal replacement for our purposes, while also making things a bit faster since we now don't need to setup a new watch. This unbreaks parallel runs of the kdirwatch test. Since the test gets compiled for each internal backend we cannot do stuff on shared paths (such as cd'ing home and watching it - we'd run into reentrancy problems) M +2 -24 autotests/kdirwatch_unittest.cpp https://invent.kde.org/frameworks/kcoreaddons/commit/a38e639ad913d1dce1f808029333fc00995b641e -- You are receiving this mail because: You are watching all bug changes.
