On 06/11/2010 16:47, Christoph Feck wrote: >> # The "Fix my mess" plan >> [... long plan involving reverting ...] >> This should reduce the risk of build failures since any application I >> missed which switched to showHide*() methods will continue to build but >> get warnings for one week. Does this make sense? > > It would be simpler to just revert the kdebase/apps commits, too, instead of > API juggling. I compile trunk daily, so if there is commit missed, I can spot > it and revert :)
Actually what I meant was to revert both changes in kdelibs and kdebase but keep the deprecated flag on the methods which would be removed for a week, to avoid breaking other applications. Might not be worth it indeed. >> Applications which have "[ ] Show <ui item>" menu >> items should not change them to "[x] Hide <ui item>" when it is >> toggled. This is clearly wrong. > > Exactly. Left-over from the days where KToggleAction did not force a check > box. > > David Faure is the maintainer of XMLGUI stuff and actually was in favor of > those API changes, so wait for his decision. OK, makes sense. Aurélien
