* Allow Lua action hooks to specify custom functions for more behavior
...like gravity, and update all current mods to match.
Spin jump and wall slide from the Extended Moveset mod now have gravity
code basically matching the original mod.
Currently, any place you'd want to use one of these new action hooks
still requires an O(n) check through all action hook per call. This
should probably be changed.
* Fix some remaining issues with Extended Moveset Lua port
- Remove divergent spin jump code
- Remove divergent roll code
- Remap roll button to Y
- Reimplement dive slide to make dive hop work like the original
- Allow spin from double jump, backflip, side flip
* Fix more issues with Extended Moveset Lua port
- Reimplement all users of update_walking_speed to incorporate the Tight
Controls edits and modified speed caps
- Fix instances of angle arithmetic to wrap properly across the mod
* Don't chop off group bits of custom action flags; assign missing groups in mods
This fixes the Extended Moveset's underwater actions. Chopping off those
bits was making the game consider the underwater actions to be a part of
the Stationary group, which caused `check_common_stationary_cancels`,
which upwarps Mario to the surface.
* Tweak roll sliding angle tendency
Rolling will now gradually (but fairly quckly) try to tend Mario's
facing angle down the slope.
This is cleaner than my old method that tries to flip Mario's angle
(wrongly) when he begins moving downward, having that logic coexist with
the logic for normal sliding actions that can also tend Mario to face
backward down the slope. Just looks ugly now by comparison.
* Disallow spin jump on slides in Extended Moveset port
This matches the original mod
* Extended Moveset: Crazy Box Bounce check
* Extended Moveset: Fix hugging the wall when spin jumping after wall kick
* Extended Moveset: Fix ledge drop snapping up to ground
Just reimplement `act_air_hit_wall` ourselves.
* Extended Moveset: Add Kaze's walking speed fix
* smlua_hooks.c: Restore option to use old API for hook_mario_action
The intent is to allow mods outside of this repo to continue working.
Co-authored-by: djoslin0 <djoslin0@users.noreply.github.com>
* Support for more granular player colors
You can now configure RGB values for shirt, pants, gloves, and shoes.
Due to some limitations, configuring shoes does nothing at the moment.
* Remove paletteIndex and friends
Restructured and filled in some remaining code to account for that.
* Add Edit Palette panel to Player panel
* Change PlayerPalette contents to an enum-indexed array, remove shoes
This gets rid of all the hokey code doing switch cases on the
different parts.
* Fix goof with player model selection box
Should actually have affect now even if a custom palette is being used.
* Fix gap in player color display list commands
The extra space was leftover from when I was trying to get shoes
working. Forgot to clean it up.
* Standardize PlayerParts enum, including for lua constants autogen
* djui_panel_player.c: Properly hook sending palette changes on unpause
Editing the palette and then unpausing should send out the packet to
everyone with the new palette changes (and update the palette preset
selection box), but since we weren't hooking that situation before, it
would stay changed only for you. You would have had to press the Back
button for it to work right.
* Allow Lua mods to continue using `paletteIndex`, `overridePaletteIndex`
This lets mod code like this still work unchanged:
if s.team == 2 then
np.overridePaletteIndex = 7
elseif s.team == 1 then
np.overridePaletteIndex = 15
else
np.overridePaletteIndex = np.paletteIndex
end
It's essentially faked, and would work strangely if the value of either
variable was inspected more closely directly. This should at least
handle the typical use case, though.
Every frame, `overridePaletteIndex` is checked to see if it was modified
from its previous value. If so, `overridePalette` is set to the preset
corresponding to the index. `paletteIndex` contains a special value that
when used to assign to `overridePaletteIndex`, it copies `palette` into
`overridePalette` to restore the real colors, which of course may not
follow the presets at all.
* characters.h: Pack `PlayerPalette` to eliminate size differences between computers
* mario_misc.c: Remove remaining "TODO GAG"
* First attempt
First attempt at having other Mario objects create ripples.
Works very inconsistently, the position of players not currently interacting with the painting seems to influence whether the ripples will work or not
* Moved nearest mario code from geo_painting_update
Instead of using the nearest mario object for geo_painting_update, I use it in painting.c functions that called for gPaintingMarioXYZPos, substituting the nearest Mario object's position instead, as well as using it for the last floor mario was over.
This method still isn't quite perfect and there's a few more things I'd like to fix/change.
1. If a Mario is closer to the painting (perhaps right against the middle) other Mario's can't trigger ripplings at all unless they get closer than the closest Mario
2. I would like for all Entry and Exit ripples to interrupt each other, meaning if multiple Marios jump in, each one triggers a ripple one after the other
3. I want to hear the other player's exit sounds
* Add currPositionValid
Sets currPositionValid to false for NetworkPlayers who are connected, who changed levels to a level different from the local player, and who are not the local player when the local player changes levels.
* Set currPositionValid
Sets currPositionValid to false for update packets that don't match local area.
Sets it to true for update packets that do match local area.
* Send change area/level immediately
Adds a function network_on_loaded_area2() which only sends a change level or area event and doesn't clear the sync ids.
Calls this function almost immediately as opposed to waiting for LOADING_LEVEL_THRESHOLD frames
* Add currPositionValid to NetworkPlayer
currPositionValid (which is used by painting.c to determine if a player should be considered for the closest player to the painting) is a variable that tells you if a network player's marioObj has a position which can be relied upon as a valid position for that player.
network_player_update_course_level sets it to false for a network player if the new location does not match the local player's location.
network_receive_player sets it to false if a network player's update is from a different location than the local player, but sets it to TRUE if the update is from the same location.
update_course does NOT set it to true, because the location is not sure to be valid until we've received a location from the player with the right location attached to it.
* Use currValidPosition correctly
Uses currValidPosition correctly, also add debug text for seeing closest player to painting.
* fix merge
* Fix merge
* Verboser logging and invpos on enter
* Multi collision arrays
* Removed nearest code, added comments
* Added simultaneous ripples on one painting
* Added commented Gamepad from Background line
* Removed debug logging
* renamed secondary network_on_loaded function
* Added early multiple paintings rippling at once
* Added multiple paintings rippling simultaneously
Co-authored-by: djoslin0 <djoslin0@users.noreply.github.com>
Added two options, one which lets user choose which gamepad to use, and one
which allows game to be played with gamepad while in the background.
These two features together allow for "splitscreen" (sort of) multiplayer by
starting up instances and joining them together and selecting different
gamepads for each one.
Gamepad choices are 0-7. If a number is chosen that does not correspond to a
gamepad, the previous gamepad (last number you were on this session that had
a working gamepad) is used.
