[PATCH] missing control for deactivating guard order in mixed selection

[Kalle Richter]

After ordering a selection x of units to guard a unit y it's not possible to undo this order if a selection x' (with x being a subset of x') is selected. Units refuse orders and return to guard the unit and commit suicide.

experienced with d884ffa0b1c763742dae483a474c327d06ff4d7a

[Stan]

Could you attach a system_info.txt, and a minimal testcase ? (Small map with few objects)

[svott]

I just add this.RemoveGuard() to the orders. The number of adds doesn't look nice, but tell me if there is a better solution I am also not sure if I missed something or added to many remove-calls.

[Imarok]

(If you want your patch reviewed you have to add the "review" keyword)

[fatherbushido]

(It seems you forget to remove the warning L4914.)

[svott]

I think I mentioned that in IRC and it is not enough to spend time for uploading another version

[sanderd17]

I think it should be possible with altering fewer lines. Did you try adding a RemoveGuard call to the "leave" functions of the INDIVIDUAL.GUARD state?

[svott]

I tried and it does not work that way because a guarding unit always leaves the guard state when the unit stops - but the guarded attribute is still set

[fatherbushido]

I don't understand the issue of the ticket. Guard order could/should be explicitly cancel in the gui. The issue would be better described as elexis reported: When I select one guarding and one non-guarding unit, I cant remove the guard as the icon is hidden if the garding units is not the first selected. According to some discussions on irc, it seems that it's not an UnitAI matter but a general gui problem, where possible orders are often based on the first unit of the selection.

I don't know if we want what the patch does. For example, if we order a unit to guard another, a walk order should not cancel the guard order, when the walk is done, the unit must return to guard.

[elexis]

Indeed not a simulation but GUI issue. Closing a duplicate of #3823.