Caranthar

Timeline Bugs: Changing Tweening Mode: When choosing a custom easing curve, hitting ENTER will close the selection window and not apply the chosen easing curve to the keyframe (probably ENTER triggers the "cancel"-button in that window instead of the "apply"-button. (Might be desired behaviour, but IMHO very unintuitive.) Editor Bugs: Sub-Entities and RMB: Clicking a sub-entity with the RMB to call the context menu doesn't work. Instead, the sub-entity has to be selected and then void space has to be clicked with RMB to call the context menu. (Might be desired behaviour, but IMHO very unintuitive. Simply clicking with the RMB could still summon the context menu while clicking, holding and dragging the cursor down could "activate" the sub-entity's timeline gizmo.) CTRL-copy and panning: When panning around the scene using MMB while CTRL is held, any items selected at that time are duplicated in place, even when the cursor is not hovering over them. (Highly annoying/confusing bug. I often use CTRL+Mousewheel to zoom faster and pan with the MMB. When accidentally CTRL is still held while panning, stuff gets duplicated left and right.) Deleting frame 0: When the first keyframe at frame 0 of an animation is deleted (even with no sprites in the animation) and then a sprite is added to the animation, Spriter crashes. Also, if sprites are already in the scene and this frame gets deleted, Spriter crashes the next time a sprite is drawn into the animation. Keyboard focus in text-boxes: Clicking the different text-boxes around the timeline is very inconsistent: play speed: Clicking the number field of the play speed will just turn this button into a text field neither putting the cursor into it nor selecting the text to be overwritten. The user has to manually click into the text field and/or select the old value to overwrite. Clicking anywhere (e.g. on the canvas) will not set the number back to the "button" state, but will remain in "text field" state. Only once the cursor has been placed into the text field, clicking anywhere or ENTER resets the button to the button state. ESC does not escape from the text field. name: Clicking the name field will turn this button into a text field, placing the cursor into the text field and immediately selecting all the text to be overwritten - Keyboard input can be done immediately. Clicking anywhere (e.g. on the canvas) will set the name field back to the "button" state. ESC does not escape from the text field. current time: Clicking the current time field will turn this button into a text selecting the text to be overwritten (but not putting the cursor in the field). Even though the highlighted content suggests that you can already enter a value, the user still has to manually click into the text field and/or select the old value to overwrite. Clicking anywhere (e.g. on the canvas) will not set the number back to the "button" state, but will remain in "text field" state. Only once the cursor has been placed into the text field, clicking anywhere or ENTER resets the button to the button state. ESC does not escape from the text field. length: Clicking the length field will turn this button into a text field, placing the cursor into the text field and immediately selecting all the text to be overwritten - Keyboard input can be done immediately. Clicking anywhere (e.g. on the canvas) does not set the length field back to the "button" state, but the appearing "apply" or "cancel" buttons have to be used. ESC does not escape from the text field. Overall, I find these inconsistencies to be highly annoying and constantly find myself jumping through keyframes or locking/unlocking sprites and bones when actually I wanted to enter a numerical value into one of the fields that suggested I could to so. Z-Order panel bugs: Z-order and sub-entities: When using a sub-entity in an animation, whenever jumping backwards through the keyframes using button "1" or the timeline play controls (even if there is only one keyframe), the z-order-panel will show the sub-entity's internal z-order. When jumping forward using button "2" or the timeline play controls, the real animation's z-order is displayed. When using several subentities, jumping backward through keyframes will always bring up the sub-entity's internal z-order that has been added last. Jumping forward still shows the real animation's z-order. Renaming z-order-panel items: After renaming an item in the z-order panel, the other items in the panel can't be renamed (double-clicking will not activate the renaming). The only way to make them renamable again seems to be to switch to another animation (or create a new one) and then switch back. When coming back to the original animation, one more item in the z-order-panel can be renamed, then it's locked up again. Furthermore, this siwtching animations back and forth only seems to help when switching to a different animation than the one working on. Cloning an animation, switching to the clone and then back to the animation will not make the other items renamable again. (Though the more you click back and forth, the more confusing the results seem to get.) Deleting items and readding them: When deleting an item and then readding it, still a _000 is appended to the name. Similarly, if deleting that sprite and readding it, it will be called sprite_001 and so on. The counters are reset when the animation is deleted though, starting to count the items properly again. More bugs: I saw a lot more weird bugs while messing around with all the stuff, but those I could not easily reproduce (the ones above are 100% reproducible though): Items getting undeletable when they had been deleted from the first keyframe, but still appeared in later keyframes - often their loc/rot/scale would also get messed up by deleting their copy in the first keyframe. Scaling-gizmos of bones (the three white dots) not staying in place, but moving about the screen by a constant offset every time another keyframe was displayed. Same thing happening to sprites. Sometimes, all handling gizmos of sprites would not show, though the selection frame would show. Sometimes, zooming out far would solve that problem and bring them back, most often not though. Crashing when adding subentities containing subentities (but not all of them and not consistently).