On 11/16/2016 02:44 PM, Even Rouault wrote: >> Another scenario would be: >> >> When clicking, the group toggle between unchecked, partially checked, >> checked. At the same time it caches the state for all subitems to >> restore the partially checked state when switching to this. >> >> At any time, when a subitems check state is changed, all parents >> invalidate their cached states and override it with the new partial >> check state. If the last remaining unchecked item is checked (and vice >> versa) clicking the group skips the partially checked state... >> >> Maybe that would be more natural? > > So imagine you have an initial state with a group G checked and its layers A > and B checked > You click on G ==> it becomes unchecked, and A and B as well > You click on G ==> it becomes checked, and A and B as well > You click on A to unselect it ==> G becomes partially checked. And you save > current state of A and B as StatePartial > You click on G ==> it becomes checked, and A and B as well > You click on G ==> it becomes unchecked, and A and B as well > You click on G ==> it becomes partially checked, and you restore StatePartial > (ie A is unchecked and B is checked) > You check on A ==> it becomes checked and G as well (and clicking on G would > switch between all on - all off) > You uncheck A --> G becomes partially checked > You uncheck B --> G becomes unchecked (and clicking on G would switch > between > all on - all off) > > Is that what you meant ? That makes sense to me. > Besides a slightly more complicated implementation, one downside would be > that > cycling between (partially)on state and off state of a group (with different > check state of its items) would require 1.5 clicks on average (1 click to go > from unchecked to partially checked, 2 from partially checked to unchecked) > instead of 1. I should check with the users if they can afford the extra half > click :-)
Yes, that extra (half-)click is exactly the one I'm worried about when thinking of your use-case. Was just thinking loud :P > > Thanks for your suggestions. > _______________________________________________ Qgis-developer mailing list [email protected] List info: http://lists.osgeo.org/mailman/listinfo/qgis-developer Unsubscribe: http://lists.osgeo.org/mailman/listinfo/qgis-developer
