| B, R |
for mapMove this currently plays one tick earlier (two ticks at the beginning before DOM.Map starts animating the movement, then each tick at N-1 when Map is animating N, then completion one tick earlier while Map is still playing animation); either mapMove ticks should be adjusted (e.g. transitionFace with tick of N, subsequent coords change at N+1, then transitionFace before next move at N+1, coords at N+2, etc.) or DOM.Map should play them differently ⋯
|
||
| B |
got stack traces telling pos is null; shouldn't happen, research and remove && when figured out ⋯
|
||
| B |
got stack traces with null cr; remove || ⋯
|
||
| B |
case 'combatMoraleGood': // XXX should play over creature's spot at the time of transition, not current ¶
|
||
| B |
user clicked before Map has reacted to mousemove and updated mouseCell which is a problem because it means mouseSide (any others) that we need here are out of date; need to figure why there is a delay ⋯
|
||
| #ccrx | B |
got many stack traces pointing out that cr is null; not sure how that's possible that no this.get('objects') had any matching obj.party == first ⋯
|
|
| B |
79 must hide duplicate spells ⋯
|
||
| B |
as with other CSS animations based on APNG, Chrome behaves strangely: it seems to "play" APNG in background even after the animation ends, so that when the animation is restarted, it "continues" playing APNG instead of playing it from the first frame. likely need to ditch APNG and use CSS animations like doing with DEF animations - possibly putting them to the same animations.css in the databank ⋯
|
||
| B |
3 Hero will pause over invalid spot (impassable, different terrain, etc.) if features change in such a way that continued movement is impossible (movement-triggered changes - a new impassable appears, terrain changes, etc.) or if client-supplied route is suboptimal. While some cases can be potentially solved by examining the move route ahead, it still won't protect against map changes unless we somehow lock certain map regions. Perhaps if this happens hero should be transported back to the last valid spot. ⋯
|
||
| B |
3 a similar problem is that PathCost is not aware of APs, they may run out over water with Boots of Levitation or over impassable with that or Angel Wings, and hero will also stop ⋯
|
||
| #mstr | I |
rather than serializing no transitions, we should serialize those with unset collect (not yet selected); also, perhaps skipTransitions is unnecessary at all, and that all transitions should be serialized in all cases to allow the client to resume some of them (trsr) ⋯
|
|
| B |
↳ Currently fails when loading certain saved games. ⋯
|
||
| B |
This still does not work in all cases. ⋯
|
||
| B |
fix $z, it must overlay shipyard ⋯
|
||
| B |
HeroAP remains full at the end of the move because its update is called twice: first in response to GenericEncounter setting AP to 0, then during running transaction of do=heroMove that triggered the encounter ⋯
|