hein added a comment.
Good idea, but every time one of those state tracking bools get copied my mind screams "BAD PATTERN". It's not the fault of your patch but I don't like how this cache handling is written - it'd be cleaner if fillCache had a return value for "I actually did work" and if the meaning of m_allProtocolsLoaded got inverted to being a dirty bit, so fillCache() itself can take care of the cache eviction code (qDeleteAll etc.) if dirty is set. Doing this instead creates a "now every time you want to change things you have to read site first and copy some boiler plate" situation. REPOSITORY R241 KIO REVISION DETAIL https://phabricator.kde.org/D11107 To: broulik, #frameworks, hein Cc: michaelh
